Report Overview

Facebook, with its massive and engaged user base, gives travel companies the ability to both build brand awareness and target consumers at a granular level to drive bookings. Conversion should continue to improve with the new dynamic ads for travel retargeting consumers across devices. Facebook Messenger and WhatsApp dominate the messaging app world with both having over one billion users. While some are excited about a messaging evolution for booking, we take a more pragmatic view with messaging being more about enhancing the customer experience. This improved experience can build brand loyalty and drive future bookings. The report discusses the best use cases across the various sub-sectors of the travel industry as well as how we see direct and indirect monetization playing out. Google leads the market for digital advertising in travel, but Facebook is making a strong push into the space. While we fully expect Facebook to grow its travel market share in a meaningful way, Google will remain the dominant force in travel digital ad spending. Travel companies would love to reduce their reliance on Google, but it is premature to think of Facebook as being able to match Google’s footprint. This is not meant to minimize the importance of Facebook to the travel industry and the financial impact of travel on Facebook. Both will become significant. Instead, we make the argument that Facebook’s gains do not have to be a meaningful headwind for Google. We envision a world where both Google and Facebook benefit from the continued move from traditional advertising formats to digital. Google will continue to be the best functional tool at driving clicks to sites while Facebook is more visual and inspirational, but lacks the clear purchase intent of a Google search. Turning to the travel industry, competition for bookings amongst online travel companies themselves and between them and the hotels remains intense. Digital advertising will remain a crucial tool to drive traffic and conversions. This is a headwind for the travel industry, where ad spending growth is trending above revenue growth, but a nice tailwind for both Google and Facebook. We believe that Facebook’s vast reach and impact on travel makes this report a valuable read for participants and investors across online travel, lodging, air travel, and tours and activities. At approximately 45,000 words, the report is equal parts financial/strategic analysis and round-table discussion with management at Facebook, Google, Kayak, Momondo, and Skyscanner, and industry experts in digital marketing, messaging and chatbots, and hospitality.

What You'll Learn From This Report

  • How Facebook has evolved from a social media company into a platform
  • Why Facebook and Instagram are a strong fit for advertising in travel
  • How dynamic ads for travel work and why they are important
  • Facebook’s advantages in cross-device tracking and conversion attribution
  • Why Facebook will not become an OTA, but metasearch could be an option in the future
  • The mechanics behind advertising on Facebook and Instagram including the types of ads, costs, and strategy that best fit the vast options available
  • Digital advertising market share estimates
  • Why ad dollars continue to move to digital and how large the opportunity is going forward
  • How much the largest online booking sites likely spend on digital advertising and what share goes to Google, and now, Facebook
  • Estimates of total travel industry spending on digital advertising
  • How much market share Facebook is likely to gain in digital travel advertising
  • The unique challenges for Facebook to win market share in travel
  • How Facebook is impacting Google, the online booking sites, hotels, airlines, and in-destination activities
  • What the future likely holds for messaging and chatbots in the travel industry
  • Why we expect messaging to be more evolutionary than revolutionary
  • An understanding of the differences between the leading messaging platforms
  • How we expect Facebook to monetize its messaging platforms
  • The expected use cases for messaging for different parts of the travel industry
  • Why travel is financially and strategically important to Facebook

Executives Interviewed

  • Christine Warner, Travel Head of Industry, Facebook
  • Neasa Costin, EMEA Travel Vertical Lead, Facebook
  • Hugo Burge, CEO, Momondo Group
  • Tore Pein Jensen, Global CMO, Momondo Group
  • Giorgos Zacharia, CTO, Kayak
  • David Low, Developer Advocate, Skyscanner
  • Colin Kim Senior, Manager Paid Growth, Skyscanner
  • Rob Torres, Industry Director Travel, Google
  • Paolo Torchio, VP Digital & E-Commerce, Two Roads Hospitality
  • Fiona Boyce, Director Social Media & Content Marketing, Two Roads Hospitality
  • Jeremy Jauncey, Founder & CEO, Beautiful Destinations
  • Conor Ryan, Co-Founder & SVP North America, SticherAds
  • Martin Soler, Managing Partner, Martin Soler Consulting
  • Frederic Gonzalo, Founder and Owner, Gonzo Marketing
  • Nick Gibbons, Co-Founder & Head of Customer Success, ReFUEL4
  • Vernon Vasu, Co-Founder & CMO, ReFUEL4
  • Volkan Çağsal, Founder & CEO, Adphorus
  • Mat Harris VP, Product, Enterprise Solutions, Sojern
  • Beerud Sheth, Founder & CEO, Gupshup
  • Preethi Kasireddy, Software & UI Engineer,

List of Figures

  • Travel Industry Facebook Page Data
  • Dynamic Ads for Travel Vs. Dynamic Ads
  • Active Users by Social Platform
  • Social Media Advertising Growth Is Maturing
  • Google and Facebook Share of Digital Advertising
  • Worldwide Digital Advertising
  • Worldwide Digital Advertising Growth Rate
  • Advertising Dollars Vs. Consumer Time Spent on Media Vertical
  • OTA Digital Advertising Spend Trends
  • Travel Industry Digital Ad Spend
  • Google's Travel Revenue
  • Facebook's Travel Revenue
  • Facebook Worldwide User and Engagement Trends
  • Facebook Active Users by Region
  • “Hotels in NYC” Google Search Test Results


The following report explores how Facebook is impacting the travel industry. We discuss the mechanics and strategy of advertising on Facebook’s platforms and dig into the new dynamic ads for travel. We will then offer an overview of the digital advertising market where we show how and why Facebook has dominated the social media space while Google has done the same in search. As advertising dollars continue to move from television, print, and radio to digital, both Facebook and Google are well positioned to take advantage of this structural change.

After walking through broader digital advertising trends, we will move on to the travel vertical specifically, where the largest online travel companies spend billions each year on digital advertising with the lion’s share going to Google. The report provides estimates on how much Priceline, Expedia, TripAdvisor, and Ctrip spend on digital advertising, what market share that likely represents within the travel industry, how much travel spends on Google and Facebook, and what the future looks like.

We analyze the revenue and value per ad implications from travel on Facebook. With ad loads now mature, the value per ad for consumers driven by views and conversion becomes even more important for Facebook’s growth trajectory. We discuss why the travel industry will be a valuable source of increasing value per ad and revenue growth. Further, Facebook’s cost per ad is much lower outside of North America, giving travel companies a low-cost way to test new and growing markets.

Facebook’s competitive advantages include the ability to retarget across devices and do so at the identification rather than cookie level, the aspirational and visual nature of the ads, it being in-app, and the fact that Facebook has a large and engaged user base. Several challenges to grow its travel footprint are that the broader travel industry is typically slow to embrace change, attributing ad performance is more difficult in a multi-touch world, Google’s continued dominance commanding the bulk of ad spend, and how complicated it is to optimize ad spend. Overall, the advantages far outweigh the challenges where we see Facebook flourishing in travel.

The report analyzes the unique ways in which Facebook is impacting the various sub-sectors in travel including online travel companies, hotels, airlines, and in-destination offerings like tours and activities.

Facebook has the two largest messaging platforms in the world with Facebook Messenger and WhatsApp. We walk through the differences between those apps and peer platforms looking at things like geographic reach, active users, technology and functionality, and use cases. We discuss why messaging will not be a major booking tool, but will become important for customer service and engagement, helping drive more brand loyalty and future bookings. Lastly, the report discusses the monetization of messaging.



Facebook has continued to reinvent itself while staying true to its core principles[1]. The overarching theme that the company laid out was:

“We are building Facebook to make the world more open and transparent, which we believe will create greater understanding and connection. Facebook promotes openness and transparency by giving individuals greater power to share and connect…”


After the company had its initial public offering in May 2012, there was tremendous skepticism as to whether Facebook could ever monetize mobile to the extent it did with desktop. Management saw that the world was becoming more mobile and Facebook pivoted to a mobile-first model where the site and app were designed with how it looks on mobile first. Revenue grew rapidly and followed consumer usage patterns; Facebook began to place ads on mobile devices. Mobile revenue went from non-existent in 2012 to over 80 percent of total revenue in 2016[2].


Today, Facebook is far more than a social network. It has become a platform that includes Facebook, Instagram, Facebook Messenger, WhatsApp, and Oculus. People still use the platform to connect with friends, but more and more, the sites are being used to conduct transactions directly and indirectly. Messenger and WhatsApp both have over one billion users that utilize the chat platforms, which are now both apps outside of the original Facebook app. Oculus, despite its $3 billion[3] price tag is more in the novelty phase, but we expect usage cases to rise over time.


Similar to how we stood at a crossroad for mobile back in 2012, it seems like that is the case in the travel business today. Conceptually, it is clear that visual ads that can be targeted at a very specific portion of the population should be effective for the travel industry. The ability to retarget a person with lodging options who recently searched a travel site for flights to Rome in October should facilitate high conversion on ads and better return on investment for the clients. We have seen broader acceptance of Facebook as an effective ad platform, but the travel industry tends to move cautiously in approaching change.


By having an ad format specifically designed for the travel industry, we expect to see adoption increase. Bidding and ad optimization are complex so we see many companies utilizing third parties to run the campaigns, build software, or offer consulting services. The largest online travel companies possess incredible resources on the technology and personnel side and are more likely to bring the ad optimization process in-house.


Google’s advertising platform is highly effective, but more pragmatic than aspirational. Travel companies need to be present in search results to drive traffic to their own sites, but there is nothing inspirational in the way that Facebook offers. This does not diminish Google’s role nor its importance in driving traffic, but simply highlights a key difference for Facebook. That difference makes Facebook a complementary advertising tool to Google. Of course, both compete for ad dollars, but it’s not an apples-to-apples comparison.

Facebook’s ability to retarget across devices is crucial for the travel industry where a person may search on mobile, but then book later on a laptop or desktop. Facebook has an advantage here where the tracking is done at the personal identification level rather than through cookies that approximate an identity. Facebook is powerful enough to track both in-app and browser behavior.


We walk through our findings in great detail later in the report, but broadly, Facebook should see its share of travel dollars in digital advertising accelerate from the low single digits to over 10 percent in the next few years. Across the board, online travel companies and hotels have told us that spending is increasing for Facebook. The gains in travel will be meaningful for Facebook’s growth path, but the headwind for Google will be quite manageable.


Expedia has noted that we could see ad spend on Facebook reach “Google-like” stature. We believe that this statement is more negotiating tactic than reality. Google remains incredibly important to drive traffic to the online booking sites. It is much more likely that Facebook takes perhaps a few percentage points of share spend from Google and more from other smaller digital platforms, television, and print.


Part of Facebook’s growth story in the past was the acceleration of the ad load or the percentage of ads a user sees on her wall versus organic posts. Facebook management has commented that ad loads are essentially mature. Rather than push more ad dollars through supply growth and risk hurting engagement, management is smartly choosing to grow more through demand and value. The travel industry will pay more per ad (especially if conversion is strong) than smaller verticals, so moving ads shown into the travel industry can help increase the cost per ad for Facebook.


It is clear that messaging will become increasingly important for the travel industry. There is debate as to whether it is revolutionary or more evolutionary. Our view continues to be that messaging will first be used to handle customer service with bots moving toward routine inquiries. We see this freeing up humans to handle more complex inquiries. Over time, we see machine learning accelerating and the type of queries being handled by bots becoming more complex.

Bots and messaging are part of the booking process, but the experience is not that robust. Additionally, training consumers to use messaging to book is difficult and largely not needed. We see mobile continuing to be crucial in the travel industry for booking and customer engagement and messaging being used to partner there for communication.

Facebook has started experimenting with advertisements here, but this is in the very early stages. Messenger will be the primary platform for advertising versus WhatsApp as it is less utilitarian. In both cases, the ad loads will likely be very small as this ramps up. The indirect monetization will likely come from the data on the messaging platforms being used to retarget consumers on Facebook and Instagram. The company is also likely to take a fee on transactions based on a percentage of the payment. For travel, this will be less important than it is in more commoditized verticals like retail. We do not expect to see usage fees given Facebook’s continued push to connect the world.


Similar to how some have postulated that Facebook has the bandwidth to compete with Google on search, others suggest that Facebook will eventually become an OTA. As is our view with Google, we believe that will not happen. The OTAs own the customer and provide customer service. This sounds simple, but with global operations and vast inventory, a lot of scale is needed. Facebook and Google could do this, but they are better served focusing on what they do well or innovating industries (i.e. Google cloud, Facebook Messaging, VR, etc.) rather than trying to duplicate what a company like Priceline has already done.


Looking out into the next decade, it is conceivable that Facebook experiments with an instant booking model or search function where it does not own the customer. Metasearch requires technology, inventory, branding and marketing, and a strong user experience. Facebook has the budget and people to do this (and why Google has done so). We do not expect this to occur, but it is a realistic long-term option for Facebook. Our most likely case is that Facebook continues to grow its ad revenue from the travel industry.


We believe that Facebook will become an increasingly important part of the travel ecosystem. Online travel companies have been early to recognize this. Expedia, for example, has publicly stated that it sees Facebook as potentially achieving Google-like levels. We do not believe it gets to that level given Google’s importance in driving traffic, but we certainly see rapid growth at Facebook’s travel business. To the extent that messaging gets monetized, that further increases the share of travel dollars Facebook is likely to receive.

Based on public comments and our conversations with individual companies, we believe online travel companies are rationally increasing spending on Facebook. These companies should not drastically move away from Google given the effectiveness there, but rather pull advertising away from other channels within digital and outside of digital. We believe in a multi-pronged approach with Google acting as a broad tool for acquiring customers and Facebook for inspiring them.

On the lodging side, hotels should create tailored advertising with a “social media first” mindset. These should be aspirational and use imagery to tell a story. Most ads on Facebook are viewed without sound and for a short time period, so simply repackaging TV ads does not work. By using retargeting aimed at small subsets of the population, hotels can keep costs per action (Facebook has many types of clicks) down and increase return on investment.


Facebook’s ads are in the news feed along with the core content, so they are more organic and less invasive than traditional methods of advertising. From a Facebook ad, people can perform tasks like getting directions to a store, downloading an app, viewing videos, adding an item to a shopping cart, or taking another action on a client’s website. Unlike traditional TV and print advertising, Facebook gives clear data back to advertisers directly and continuously.


According to Facebook, there were over 60 million active business pages based on Q2’2016 internal data. This is the simplest way to advertise on Facebook where all a business needs to create a page is a name for the page, an about section, and profile and cover photos. Users can easily add a call-to-action button at the top of the page that directs a viewer to do something like visit a website. The pages are free, but promoting them effectively costs money.

The chart below shows the likes and followers for leading participants in the travel industry. Anyone that likes something automatically follows, unless the user chooses to unfollow. That would be the case where they want to no longer see posts, but want to remain friends. The Facebook pages help for brand awareness and provide the travel companies with a list of people who are already likely to be interested in booking through that provider.

Source: Facebook, as of December 2016



Photo ads are the most straightforward method of advertising on the platform. Below is a example where the call to action is the book now button. Broadly speaking, photo ads should be visually stimulating, high resolution, and simple to understand. Facebook recommends at most 20 percent of pixels dedicated to text versus the image.


With users watching over 100 million hours of video each day on Facebook, this has become an increasingly important format. According to independent Fors-Marsh tests (an applied research company that helps organizations and U.S. government agencies make research-backed decisions), people can recall mobile news feed content after seeing it for only a quarter of a second. The value of a video ad happens quickly and increases with duration.

Nielsen found that up to 47 percent of the value in a video campaign was delivered in the first three seconds and 74 percent of the value was delivered in the first ten. Based on Facebook’s internal data, the company also found that 65 percent of people who watch the first three seconds of a video will watch for at least 10 seconds and 45 percent go on to watch for 30 seconds.

Longer engagement drives results. After three seconds, 32 percent of the cumulative impact to brand awareness and 44 percent of purchase intent is realized; these numbers jump to 65 and 72 percent after 10 seconds.

Source: Facebook and Nielsen

Ads should be designed to be effective without sound. Facebook found that if sound starts playing unexpectedly, 80 percent of users react negatively. Instead, people prefer to click to opt in for sound if they are interested in watching the video. Adding captions to videos tends to increase view time by 12 percent on average.

Overall, the most successful video ads are visually stimulating, simple in the message, engaging, and designed with an opt-in for sound in mind.


An advertiser can show users up to 10 images or videos in a single advertisement. This ad format can be used to show multiple products, different features of one product, to tell a story, to explain a process, or to create a larger canvas with multiple images.


This ad format was created to provide what amounts to video-light ads in emerging markets with lower bandwidths and connection speeds than developed markets. Facebook found that a 15-second slideshow would be a 5x smaller file size than a video of the same length. The advertiser can choose still photos or an existing video to capture three to seven images and put them in the slide show. The user then chooses image duration, image ratio, transition ratio, audio, and thumbnail along with being allowed to use text overlays.


While slideshow is a more basic technology targeted at emerging markets, canvas ads are focused on more developed ones with mobile in mind, specifically for iOS and Android. Ads use a combination of full-screen videos and still images, text, and call-to-action buttons. Users can swipe through, tilt, and zoom to view panoramic images and zoom in to view images in detail. Canvas loads quickly, as much as 10 times faster than the standard mobile web.


Advertisers are able to automatically market every one of their products with Facebook’s technology facilitating the targeting. A business can show single or multiple products to customers who have visited their website or app without configuring thousands of ads. The first step is to upload the product catalog to Facebook Product Manager and the list of products to be advertised along with the info in the ads.  Next, a pixel on the website or app events on an app is used to show when products are being viewed, added to a cart, and purchased. By tracking behavior, an advertiser can see if a customer came close to purchasing, but did not do so and offer a deal on that product or a similar one. Another option is marketing complementary products after another is purchased. Facebook provides reports on which ads are doing well so clients can optimize marketing spend better. The ads are always on and work across devices.


Dynamic Ads for Travel became officially available on Facebook and Instagram in June 2016. In a recent Skift article,[4] Facebook told Skift that Dynamic Ads for Travel’s destination catalog is designed for travel brands seeking to display images of destinations in their ads. The chart below shows some of the differences between dynamic ads as a whole and the travel focused ones. The main difference is the ability to target the audience with destination- and date-based products.


The ads are currently designed for hoteliers, but will be expanding to the other travel verticals. A hotel company can set up one campaign to promote all of its inventory. The campaigns automatically deliver ads with real-time pricing and availability to a targeted group of people that have expressed an intent to travel to a given location and in a given time period. By using the platform, hotels can sell lodging to people who have searched or booked flights as well as consumers who visited the hotel’s site. The ads will appear across devices regardless of what device they used to view the product initially (more on this later in the report).

Bidding is CPC, CPM, or OCPM.


Sojern is travel’s leading data-driven performance marketing engine. Currently one of the fastest growing travel tech companies, Sojern works with top travel brands and independent hotels in North America, Latin America, Europe, the Middle East and Africa, and Asia-Pacific. The company has its headquarters in San Francisco with key offices in Dubai, London, New York, Omaha, and Singapore.

Mat Harris is a startup veteran with extensive product and commercial leadership experience in mobile advertising. After building and managing a successful digital agency and a video advertising platform in Austin, Texas, he moved to San Francisco to lead product management for InMobi’s rich media authoring platform and was the head of advertising products at Tapjoy. Mat frequently speaks on topics including dynamic personalization, mobile advertising, marketing strategy, leveraging big audience data, and the future of advertising.

Skift: Can you provide an overview of Sojern at a high level and how it’s uniquely positioned with the sole focus on travel?

Mat Harris: I would say that Sojern considers itself a performance marketing engine. We focus only on the travel vertical. We believe that our unique focus gives us the ability to develop a lot of specific intelligence and also expertise that’s focused around the travel vertical. What we’ve seen is that the verticalization of advertising allowed us to specialize, to focus on developing a core competence around the things that matter for travel advertisers specifically. That what we’ve done here at Sojern.

Skift:   In general, how is data used to optimize ad campaigns?

Mat Harris: We develop and craft audiences that are specific to verticals within travel, so hotels, airlines, car rental agencies, destinations, they have different goals and different data they’re trying to accomplish with their campaigns. We craft audiences that are specific to accomplishing the goals of the advertiser and those are further refined based upon the vertical that they’re in.

We have a large number of data partnerships with sites and apps in travel that give us the ability to see searches, see bookings for flights, hotels, car rentals, and also searches for destination. This gives us the ability to identify the right audiences to serve ads to, as well as form our optimization algorithms to make sure that we’re driving the best results that we possibly can and improving the quality of those audiences over the course of the campaign.

Skift: Do online travel companies and hotels utilize Sojern and Facebook differently?

Mat Harris: We work with both. The primary reason that the advertisers work with us is because we bring data to bear that they can’t find anywhere else. We bring the data and run those campaigns for them in a managed service. On Facebook, it’s truly retargeting.

Facebook had opened up its ad platform via the Facebook exchange and then recently, like in September, shut down the Facebook exchange. Via the Facebook exchange, you could access Facebook inventory very much like accessing any other ad exchange. You could use your own audiences, you could target inside of Facebook leveraging what you know and you could optimize those placements just like you would anywhere else and it would be exactly the same as running on any other exchange.

Facebook exchange was limited only to desktop and as Facebook users were migrating more and more toward mobile, Facebook exchange advertisers were losing access to audiences. Facebook used that opportunity to transition everybody into now advertising on Facebook only via the Facebook marketing API and closing the exchange. Part of that transition was also the release of a new product which they called product ads some time ago.

Dynamic Product Ads (DPAs) were created essentially for retail, so for the likes of Home Depot or Amazon, so they could leverage product ads in order to have a real time retargeting product inside of Facebook. If you happen to be shopping for a hammer on Home Depot for example, or a water heater might be a better example, and then you go to Facebook you could be retargeted by Home Depot for that water heater you just looked at because of dynamic product ads. What Facebook learned is that as they released dynamic product ads, the travel advertisers were adopting DPA even though that DPA product wasn’t built for them. DPA product had in its product feed requirements such as a field for the condition of a product, whether it was used or new. This as a required field would still be there even for travel advertisers. They would just mark something used or new even if it was a hotel room or a destination or something like that.

What they realized was that travel advertisers were going out of their way to try to use a product that wasn’t built for them because they believed it solved a need for them to be able to provide some real-time ads that were personalized on Facebook. Facebook then created dynamic ads for travel several months later. We were able to bring some of the very first advertisers across into dynamic ads for travel.

Skift: How did Facebook change the ads for travel?

Mat Harris: Facebook updated the product feed and the pixel parameters of dynamic product ads in order to make them more relevant for travel advertisers. There’s now a feed format that’s specific to hotel advertisers that’s relevant to hotel descriptions. Also, real-time pricing and availability can be brought into these campaigns. You’ll begin to start seeing ads on Facebook that actually don’t just show a base price for a hotel room, but also the real-time price and take availability into account so they won’t show you a room that isn’t available anymore. That’s something that is specific to hotels.

They’re soon releasing the air and destinations version of this as well. Increasingly, you’ll start seeing airlines and destinations also leveraging them with specific feed formats that are customized for those uses. Now you’re moving from a product that only worked for hotels, to one that now also works for OTAs and also works the metasearch sites, it works for airlines and it works for destinations.

Skift: What are the limitations on cross-device tracking?

Mat Harris: Well I’d say the only limitation is now you are required to leverage or to trust Facebook’s attribution. You don’t have the ability to bring in third party attribution or any of their vendors to verify. You’re trusting that Facebook, the trust is well placed by the way, I’m just sort of underlining one of the challenges or the issues. As you run a campaign on Facebook, you’re inherently trusting that Facebook is going to use their opted in, logged in behavior to be able to retarget a user no matter what device they happen to be on. They visited your website on a desktop but now they’re on Facebook on mobile, you’re going to serve them the ad in mobile based upon what they searched on their desktop. That’s going to be facilitated because that user has logged in on both of those devices and Facebook has a deterministic connection where they have the ability now to retarget that person on those devices.

The effectiveness there, I think it’s pretty much unquestionable that that is a very effective way to retarget and also a very trustworthy attribution method.

Skift: Does tracking require a login to Facebook at the site level or can it just be in the background, saved in the browser?

Mat Harris: The pixel is always at the browser level. If you’ve ever, on your mobile device, ever logged into Facebook in the browser then there is a connection from that browser to your Facebook account. If you have a conduct travel search on a browser where you logged into Facebook even once in a while, Facebook has that ability to link your pixel because they’re dropping a Facebook retargeting pixel on the travel site. Now they’re syncing that back to your Facebook ID and they’re using that as the basis to retarget you inside of Facebook.

Skift: Facebook had been previously used more for brand awareness and we’re starting to move toward Facebook helping facilitate bookings directly. How much do you see that progressing where Facebook is now much more than brand awareness and is actually driving bookings that can be attributed to them?

Mat Harris: I think it’s already there. I think Facebook is in the position of considering themselves being in the performance business. They are essentially working almost exclusively on retargeting as a solution. I would say that they’re very focused on driving direct bookings and being able to prove, via their attribution, that they’re actually driving attributable conversions back to the advertising.

Skift: You mentioned that Facebook is going to be rolling out the format specific for destinations and airlines. It already has the hotels. For you guys and maybe broader for the advertising performance market within travel, where do you see a big chunk of that market developing? Is it going to be hotels? Are destinations going to be taking Facebook on more seriously? Where do you think that traction for dynamic ads is really going to take root?

Mat Harris: I think if you look at travel in general they already prioritize it in order that they see the market. Hotels are first, hotels spend the most money, they’re the largest vertical in travel. They have the most to gain from retargeting people on Facebook based upon their hotel searches. Being able to retarget people for a hotel based upon their flight searches is also an area that I think will start to take off and grow and provide substantial value.

You’ll see airlines adopt it, but retargeting for airlines has to be something that has happened instantaneously or very, very fast. Usually it’s an airline search that is the very first indication that somebody is starting to book their travel or starting to become in the market for travel. I think you will see airlines reap some benefits of being able to be on Facebook. They’ll bring less money in than hotels. I think you’ll see that and destinations even less. I think you’ll see that Facebook has rightly prioritized in terms of market size, in terms of business opportunity, the verticals inside of travel they want to serve first.


Based out of Istanbul with a global reach, Adphorus provides an intelligent advertising platform for marketing automation and algorithmic optimization for the travel industry. It became a part of the Facebook Marketing Partner program in 2014 and since then has been passionately trying to make advertising efficient and profitable by using technology and data. Adphorus is working with global travel brands such as Trivago, Kayak, NH Hotels, MakeMyTrip, KLM Airlines, and many others.

Volkan Çağsal, founder and CEO, previously co-founded the social media agency PUBLIK and was head of digital at Fox International Channels and Myspace.

Skift: How does your sole focus on the travel industry helps Adphorus position itself as a firm?

Volkan Çağsal:  Adphorus’ unique approach is a result of our smart algorithms that optimize

Facebook campaigns for better performing ads that ultimately deliver more bookings. This is the

reason we call ourselves “marketing science” and we help our advertisers to develop a

scientific approach to Facebook marketing with continuous optimization.

Our exclusive focus on travel makes our solutions super relevant for our travel advertisers. We are developing our platform based on the needs of travel advertisers, working closely with Facebook for new travel-specific product launches. We are hiring our team members with backgrounds in travel, and we are working on partnerships with multiple players in the travel tech and supplier ecosystem.

Skift: How is data used to best optimize advertising?

Volkan Çağsal: Our flagship product is called Marvin, an algorithmic optimization engine. Marvin is an autopilot for Facebook campaign management which handles budget allocation and bid management on behalf of the client to deliver on their target costs. Marvin uses advertising data provided by Facebook as well as client data such as revenue, which allows our clients to set a cost per booking or ROI/ROAS targets for their Facebook campaigns. They leave the rest — hitting the targets and scaling — to Marvin.

Another optimization solution we provide is the Dynamic Ads Optimizer, which places properties or destinations into value clusters based on their performance in order to optimize on property/destination level. The Dynamic Ads Optimizer helps clients to manage their portfolio, taking into account travel dynamics such as seasonality, political events, etc.

Skift: How differently do online travel agencies and hotels utilize Facebook? What are the key differences?

Volkan Çağsal:

  1. Dynamic Ads

Both sub-verticals use dynamic ads for remarketing activities on Facebook. Hotels chains are using hotel property level targeting where OTAs utilize both destination level and hotel property level targeting. Accordingly, creatives are more extensive for OTAs.

  1. Cross-Sell Opportunities

Since most of the OTAs are present in multiple verticals (flights, hotels, cars, etc.), they have significant cross-sell opportunities, which can be exploited through Facebook Dynamic Ads. The most common scenario is to capitalize on the audience of their main vertical (i.e. flights) by showing them ads of their growing business (i.e. hotels).

  1. Package Opportunities

Most of the OTAs have package deals that combines flight and hotel bookings and long weekend getaways. They promote these products in separate Facebook campaigns targeting mostly families and couples in their 30s.

  1. Campaign Targeting Structure in Remarketing

In general, OTAs have higher website traffic than hotel chains, which is reflected in their targeting segmentation in Facebook campaigns.

With larger audience size, OTAs tend to segment their target groups according to user behavior on the website, placements, age, etc.

For remarketing campaigns of hotel chains, where the audience sizes are small, over-segmentation is one of the most common mistakes. We suggest keeping the target groups as broad as possible.

  1. KPIs

As the nature of their business, we see that OTAs have stricter KPIs across all marketing channels including Facebook. Hotel chains usually have generous targets in terms of revenue and ROAS, which allows good delivery and optimization on Facebook.

Skift: What are the limitations of retargeting on Facebook and what can be improved?

Volkan Çağsal: With the recent launch of Dynamic Ads for Travel (DAT), Facebook now has a native retargeting solution for travel advertisers. DAT provides the capability of remarketing hotels, flights, or destinations with relevant information that is dynamically displayed on the ad. DAT also provides targeting options, which allows advertisers to segment audiences based on their actions, intent, booking window, recency, etc. This combination actually makes DAT one of the most powerful offerings out there.

However, we believe there are three major missing pieces on this offering right now, which we have built Adphorus products to solve. One is property/destination level optimization, which we handle with the Dynamic Ads Optimizer. The other one is the difficulty of providing creatives for each individual destination among thousands, which we solve with our Adphorus Media Library that supports 3,000 destinations. We are automatically matching photos from our library with the advertiser’s product catalogs, which allows them to advertise with relevant photos and optimize their ads. And the last one is, technical difficulties of creating product catalogs of hotels, destinations, and flights in a Facebook-compatible format, which can take months in our experience. Our Automated Travel Feed solution requires minimal technical effort on the client side and provides a seamless experience on getting the client data onto Facebook.

Skift: How do you see messaging/bots/AI impacting the travel industry?

Volkan Çağsal: Messenger bots on Facebook is a new trend that is influencing the travel industry as well as others, which we think will play an important role in the future of areas ranging from customer service to actual hotel searches. The messenger bots will also come with their own advertising environment, which will allow advertisers to promote their bots/messages. We are still in the very early days and there will be a discovery period for travel, but I believe investing early in this area will create a strategic advantage.

Skift: When you look at the traditional ad formats on Facebook, and now the Dynamic Ads for Travel, in what part of the booking process do you see Facebook gaining the most traction?

Volkan Çağsal: The size of Facebook’s user base is important because travel advertisers want to ensure that the platform can actually scale up before starting to invest in it and work toward their goals.

The other point is, Facebook now has a full suite of ad products designed for both higher funnel objectives, such as brand awareness, and lower funnel objectives, such as driving conversion via remarketing.

Remarketing is definitely the lower hanging fruit as these users have already shown intent and travel clients should not risk losing them to competition. Mobile app install and engagement ads also demonstrate solid performance. Branding/prospecting is a trickier objective for travel due to the fact that many travel advertisers are running branding campaigns with strict performance targets and currently it’s not possible to have the travel intent of the users on Facebook.

We see significant traction on mid-funnel and lower-funnel objectives with Facebook and expect an increase in higher funnel/branding and prospecting objectives, as there is still much room here for improvement.

Skift: As Facebook grows, do you see it taking travel ad dollars from Google or is it more about digital advertising spend growing faster than other platforms?

Volkan Çağsal: Travel advertisers allocate quite significant budgets to Google. They cracked the code on search marketing a long time ago as this has been their main channel for a long time, which also means they are quite saturated on this channel with their current KPIs. This also results in serious dependency and an incentive to break that dependency.

I think it is not a matter of shifting the ad dollars from Google, but trying to create a second channel that has the potential to be as big as Google both in terms of scale and performance. Scale-wise, the only possible alternative channel is Facebook in my opinion, and performance is getting better and better for all the advertisers we are working with.

Thus, this is more about Facebook travel marketing growing much faster than digital advertising spend growth.

Skift: With European lodging being more fragmented than in the U.S., how does that impact travel marketing strategy?

Volkan Çağsal: Having big hotel chains in a more consolidated market like the U.S. gives these chains the flexibility to build their teams and know-how and represent their properties directly on Facebook. In a more fragmented market like Europe, it’s difficult for the long tail hotels to build this know-how, so this leaves more space to metasearch and OTAs to represent the properties indirectly.

Skift: Google’s CPC bidding is straightforward to do, but extremely complex to manage optimally. Is Facebook similar where most smaller hotel companies hire outside parties to run the platform or is it simpler and they manage in-house?

Volkan Çağsal: Facebook recently introduced travel-specific products for advertising, and it will likely keep introducing new technologies. Most travel companies, independent of size, want to stay competitive on Facebook, and therefore need to implement new solutions as soon as possible. Running the operations via an outside party as managed service versus in-house team is more of a resource problem. If the client has the capacity to build this team and know-how in-house, this is the way they choose.

However, even if there is an in-house team, the main challenge for all advertisers is still the optimization part, managing bids and budgets for thousands of ads each day, which carries at least the same complexity (if not more) with Google bidding.

This is the main reason why Facebook Marketing Partners exists. We make the adoption seamless and help our clients achieve the best performance possible on Facebook. Campaign automation and optimization is quite time-consuming and requires domain-specific knowledge. Therefore, working with a Facebook Marketing Partner is the ideal solution for travel companies, independent of their decision regarding managed service versus in-house team.

Skift: What else should we know about the industry and Facebook’s impact?

Volkan Çağsal: It’s been a year since we announced our travel repositioning and dropped all the other verticals. We made this choice as we believe in Facebook’s potential to become a core channel for travel advertisers with their travel-specific product roadmap and scale, as well as our ability to add value and close the missing gaps with the travel-specific products we are launching at Adphorus. We still think we are in the early days, but the traction we got from the advertisers all around the world and our success stories with some of the biggest global brands showed us we are on the right path.

If we also add the exciting future from AI to messenger bots into the mix, we are looking forward to see how Facebook disrupts travel marketing.

We encourage all the travel advertisers to develop a scientific approach for Facebook marketing to optimize their investments.


Advertising on Instagram is similar to Facebook in the types of ad campaigns. Instagram ads include photo, video, and carousel. Dynamic advertising and the new travel product are also used on Instagram. The difference is more about the platform itself rather than the type of ads. Instagram has a younger demographic and is more photo-driven, so ads should be visually stimulating to engage users. With Instagram being a newer platform with lower ad loads, growth rates should be higher than on the Facebook site and app (albeit off a much lower base).

For insights into the Instagram platform, we spoke to Jeremy Jauncey of Beautiful Destinations.


Beautiful Destinations is the award-winning creative agency behind the largest travel and lifestyle community on Instagram, Facebook, and Snapchat. The company’s mission is to inspire and celebrate the beauty of the world, through visual storytelling of people, places, and experiences. Everything it does is focused on creating “social first” content — content shot specifically for social (not traditional) media and they do it for some of the world’s leading brands along with governments and NGOs. It has almost 14 million followers in 180 countries on Instagram, Facebook, and Snapchat.

The business marries art with science — content creation with visual analytics. Using convolutional neural networks and bayesian statistical modelling, it uncovers the visual fingerprint of images and video, to help clients create more engaging and better performing online advertising campaigns.

Jeremy Jauncey founded the company in 2012. He is a serial technology entrepreneur and investor passionate about travel, digital media, health, and entrepreneurship. He believes in building and investing in scalable, data-driven businesses that are fun and have a positive impact on the world.

Skift: It looks like you guys have eight million Instagram followers. How did that happen? Also, can you walk us through a high-level view of Beautiful Destinations?

Jeremy Jauncey: The business, Beautiful Destinations, is the creative agency behind the largest dedicated travel community on Instagram. We have actually just under 14 million followers in 180 countries across a portfolio of travel-orientated accounts. Beautiful Destinations is the hero account that you saw. We also have Beautiful Hotels, Beautiful Cuisines, Beautiful Matters, and those really encompass the destinations being focused on. Hotels for property, Cuisines for the food and the gastro-travel story, and then Matters for nightlife, entertainment, and culture within travel.

The basic premise is our Instagram, Facebook, and Snapchat audience turn to us every day to be inspired about where to travel, where to stay, where to eat, what are the trends that are happening across the world in terms of places to go. Really, I guess, the space that we try really hard to occupy is inspiration. Starting off the idea of where a person should travel, where they should stay, and so on.

The business model is really simple. We have one part which is our creative agency, which actually shoots the advertising campaigns for a bunch of travel brands. Our biggest audience I’d say from a client standpoint is tourism boards and hotels. What we educate them on, and what we’re really pushing harder in the travel space is this idea of creating social-first content. The idea that actually social media and digital in general should be the starting point of all advertising campaigns for travel brands, because if you have a meaningful audience online you are able to test, learn, use data, create content that can then inform the rest of your marketing strategy. That is our business model.

Brands will come to us, we’ll shoot photo campaigns, drone campaigns, video campaigns. We will then post that content across Beautiful Destinations, talking to our audience about the partner, and really helping that individual client to tell whatever story it is that they want to tell to a mobile social audience. We’ve built out the partnerships that we have. Our clients range from Marriott and Hilton through to New York City, the Philippines tourism board, the Hong Kong tourism board. These guys are now realizing that social media, and particularly Instagram, is one of the most important channels for acquiring new customers.

Skift: Could you discuss how Instagram is different than Facebook from an advertising perspective? I know demographically you tend to skew younger and there are more pictures, but given your expertise here, you probably have a lot more granularity on the big differences that you see when you run campaigns on Facebook or Instagram as well as Snapchat.

Jeremy Jauncey: Of course. It’s very interesting when you look at the differences. The way I think about them is on a spectrum. At the very raw, authentic, real-time level of the spectrum you have Snapchat. As anyone who reads the media about Snapchat, amazing business, explosive growth, skews very young, and is very focused on this idea of creating a real-time narrative. Whether it’s a travel brand or not, a real-time narrative of what’s happening in your day-to-day life.

If you take a step a little bit further away from real time, raw, authentic, you see Facebook. Facebook is a big platform for us. We have over a million fans on our Facebook page and we do lots of Facebook advertising with our clients. Facebook and Instagram we see as a much more focused advertising machine.

Where Facebook has been incredibly strong has been the development of video. There was an article, I think it was the middle of last year that came out from one of the heads of Facebook in Europe who said that the written word was dead. That was said in the context of how much video was exploding on Facebook. Not only the consumption of video by Facebook users but also the distribution, the sharing, and the engagement of video from people on the platform. We have a big video business that I can speak to in a second, and we believe video is going to be the future of all travel advertising, and Facebook sits very, very strongly there.

Then, in our opinion, at the very top of a curated, managed, advertiser funnel you have Instagram. Instagram is a celebration of beautiful pictures; it’s a celebration of visual storytelling around photographs. They’ve made some massive inroads in releasing video capabilities. They just announced the release of targeted ads for brands in Instagram stories, and the development of Instagram analytics that now allows you to see the impact of Instagram story traffic from a conversion and a lead generation standpoint. I think it’s absolutely huge, this ability now to show and create really amazing visual stories that are interactive, that are engaging on Instagram and include clickable links that push people out, whether it’s to your web page, whether it’s to the booking part of your site, whether it’s to your app. Whereas in the past, really the only path out of Instagram had been the link in your bio.

We have just finished an Instagram story for Hilton. It’s one of the clients that we’ve been creating Instagram story content for.

In a talk I did a couple of years back, I pushed on the travel industry that it was, and still is, the second-slowest industry after financial services to adopt Instagram, which is completely nonsensical given how visual the travel product is, how social it is, and how much time people spend on Facebook and Instagram engaging in travel. I was at a conference in Brussels just before the end of last year and one of the leaders of Facebook travel in Europe was giving one of the keynotes alongside me. She basically explained that the number one topic on Facebook is travel by a country mile. The opportunity then for travel brands to think of a different way that they can tell stories to people through Facebook and Instagram is absolutely massive.

Skift: In the past, Facebook has been successful at building brand awareness. With the new Dynamic Ads for Travel on both Facebook and Instagram, how difficult do you think it’ll be for bookings to actually be facilitated through Facebook directly?

Jeremy Jauncey: I think the days of Facebook and Instagram being thought of as just brand awareness are now very long gone. The digital advertising tools that you have through Facebook and Instagram now are incredibly powerful. In fact, I think Sheryl Sandberg said back in the last year, “Facebook doesn’t think of itself as a social media company. They don’t use the word social media internally. They are a media company that actually drives transactions, and drives sales.”

Brands are spending advertising dollars with Facebook are spending them to drive transactions to their businesses. I really think that we’re so, so far ahead of it just being thought of as a brand awareness play where you can’t measure the impact. Maybe we were there a couple of years back but we’re way past that now.

The second thing, which actually I think is the most important thing, is the effectiveness of Facebook and Instagram advertising is 100 percent down to how good the content is. Everyone has said for a long time that content is king, but when it comes to actual real meaningful advertising on both of these platforms, the biggest opportunity, I feel, for travel brands is to invest in creating social-first content. What that means is today the majority of travel brands are taking the advertising campaigns that they shot to run in magazines, to run on billboards, to run on TV, all the traditional advertising approaches that they’ve always done and then they are dropping them onto social media with the hope that they will perform.

What some of the more progressive ones are realizing is that if you invert that model and instead of taking content for traditional and drop it onto social and you create social-first content, you then actually post it and you start to see the results from people in real time engaging in your content on Facebook or Instagram. You can then derive the insight you need to inform the rest of your marketing spend.

To give you a real-life example, we shot the Philippines tourism board’s advertising campaign in May of last year and we created a series of video assets, drone assets, and photo assets. Over the course of the subsequent months, they posted the content on their Facebook and Instagram. We posted it on our Facebook and on Instagram on our organic channels. We started to see which content was performing the best; what was getting the most likes, what was getting the most engagement. Then with that data we came together with them and we started to buy Instagram and Facebook ads to target the different demographics and user groups that they wanted to start visiting the Philippines.

That proved very successful as we rolled out the different advertising campaigns on Facebook and Instagram using content that we knew had performed really well before we actually ran the ads.

Skift: We’ve been talking to some of the hotels and the OTAs, what we’ve been hearing is that they’re pushing the Facebook platform from basically non-existent to becoming a line item in budgets.

What percentage of advertising budgets tend to be for Google versus Facebook and everybody else? We believe that 75 to 80 percent of digital advertising tends to be Google, 20 percent everybody else, and Facebook’s probably less than five percent. I was wondering what your thoughts were on those metrics.

Jeremy Jauncey: That’s a good question. I think the things that you guys have gathered thus far are relatively representative of the travel industry today. I think that’s probably fair including our own client base. What I see changing, and what I predict will be changing in 2017, is a dramatic increase in Facebook and Instagram advertising. I think what we’re facing now is an opportunity, an enormous opportunity in travel where we’re thinking about the nature of Google advertising. Google advertising is driven by keywords. It’s driven by finding ways to tell a verbal story about what a travel experience is. Then contrast that with what you can do on Facebook and Instagram. You can use canvas ads on Facebook to create fully interactive, fully immersive visual ad units that show you and actually take you through the real experience of being on that travel journey.

When you compare the moment, the amount of money that’s going toward pushing traffic through Google versus immersive ad units that actually really engage people, maybe today they’re going more with Google, but I really see that changing. We know more and more people are spending time on Facebook and Instagram.

Skift: When you guys talk to clients, when they’re looking at their budgets, do they look at Google and Facebook as comparable or do they look at Google as a search acquisition channel and compare Facebook and Instagram to television and print? Also, it sounds like you had success with Facebook moving down that funnel.

Jeremy Jauncey: Absolutely. I think the first assessment that you made is very fair, that our clients are certainly seeing Google as a very targeted kind of performance marketing. Really being able to hone in on the traveler that is very specifically searching for the hotel in Crete from the 23rd to the 30th of August. I think where the opportunity on Facebook and Instagram lies is the ability to, in the first instance, inspire the person. Inspire them by showing a really visual and engaging ad. Now we actually really move them through the funnel, all through Instagram and Facebook. We’re seeing our clients moving more and more of their budgets toward Facebook and Instagram advertising.

Actually, toward the back-end of last year we had two big tourism boards move budget that they allocated for TV advertising toward us. This was one of the first times we’d seen that happen, which was incredibly encouraging because when you look at the opportunity from a budgetary standpoint, social media is still in many cases a free service, or a service where the budgets aren’t comparable to big advertising budgets. I think what still blows my mind is that hundreds of millions of dollars if not billions of dollars are spent all over the world on billboard advertising and magazines and traditional TV, when we know that the viewership and circulation are declining in a dramatic way.

Skift: We’re talking about ROI and this idea of social becoming much more conversion focused. Are there particular types of content that work better for conversion than other when you’re posting the different campaigns?

Jeremy Jauncey: That’s a great question. We have seen the most measurable impact on hotels. Largely because the hotel transaction is relatively easy to measure versus at a tourism board level where it can take a lot of time to measure the impact of our campaign on inbound tourism. There are tools that Facebook is building and developing that we’re working with, and our tourism board clients are working with to measure that more effectively, but actually hotel has been where we’ve seen the biggest impact.

Skift: Are you putting explicit calls to action in the content? Whether that’s an Instagram story, or probably moreso on Facebook, or are you just letting the content speak for itself?

Jeremy Jauncey: We really are. No, we are testing so many different variations of how you bring the client into the story. If you actually look at Beautiful Destinations, if you look at our feed, the thing that we pride ourselves on and that we’ve always prided ourselves on is that you would never feel like you were being advertised to. We will never take a client’s marketing material and push it on our site. We will create the content ourselves and then we will weave them into the story.


Advertisers on Facebook have different goals, but essentially it all boils down to brand awareness and call to action. Such actions include app engagement, app installs, clicks to website, event response, lead generation, local awareness, offer claims, page likes, page post engagement, store visits, video views, and website conversions.


Facebook’s advertising platform is based on an auction process. Facebook assigns a “total value” based on the amount bid, how likely it is that showing the ad to a person will lead to the desired outcome, the quality of the ad, and how relevant the ad is to the target customer. The ad with the highest total value wins the auction. These auctions are continuous where a winner is determined every time a person is eligible to view an ad. The participants in each auction would be advertisers targeting an audience that the potential viewer falls into.

This is a bit complicated compared to traditional ad buying, but rewards advertisers for creating better ads.

Facebook calculates bids with the goal of spending an advertiser’s entire budget and getting the most value from the ad portfolio. The company recommends automatic bidding for advertisers that do not have a specific value in mind for the result they are trying to optimize or if a buyer is still trying to figure out the right manual bid.

A manual bid specifies the maximum an advertiser will pay for the result an ad set is designed for.

Quality by its nature is qualitative, but Facebook uses its technology to rank ads in a given auction based on a quality score and the amount bid. The ranking is essentially a score of how interested Facebook thinks a target customer will be in seeing the ad.

Feedback impacts quality where poor performers tend to be ads, where people hide or choose not to see the ad. Metrics tracked include things like clicks, app installs, video views, and users clicking “I don’t want to see this” on an ad. Higher positive feedback leads to a better relevance score, which became visible to advertisers in 2015. The score ranks an ad from one to 10, with 10 meaning the ad is highly relevant. The scores are relevant for call to action type ads, but not brand awareness. A higher score means an ad is more likely to succeed and will be cheaper for the advertiser since bids are won based on price and quality.

One key to a higher relevance score is targeting an appropriate audience. The wider the audience, the less relevant an ad is likely to be. Additionally, the more vivid and noticeable an ad is, the higher the score will be.

One thing to note is that a high relevance score is not the goal, but just a useful metric. Things like conversions, app installs, clicks to a website, and purchases are what need to actually be tracked by advertisers. Higher scores tend to lead to better results and make the bids effectively cheaper.


Facebook has a wide array of advertising campaigns, leading to vast pricing options.

The most well-known formats would be CPC, CPM, OCPM, and CPA.

The amount an advertiser is charged each time someone clicks on the link in an ad to a website or app store. The total charges are based on the amount the advertiser spent on the ad divided the number of link clicks the ad received. This is useful for call-to-action type ads.

CPM is cost per thousand impressions (how many times the ad is viewed). This format is appropriate for brand awareness.

An advertiser provides Facebook with a maximum cost to achieve a goal. According to AdEspresso[5], campaigns can be maximized for website conversions, website clicks, engagement, reach, page likes, mobile app installs, and others. It is a very effective, though costly approach where OCPM will be higher than traditional CPM, but so will conversions.

The process is similar to OCPM. Instead of paying per 1,000 actions, the advertiser specifies a maximum amount to pay for each specific action.

A sampling of the multiple other pricing options is below:

Advertisers pay based on the number of comments.

The reaction button on an ad allows people to share different reactions including: like, love, haha, wow, sad, or angry.

Viewers can share ads or posts on their own or friends’ timelines, in groups and on their own pages.

Engagement is the total number of actions taken.

A check-in occurs when a viewer updates their status in their Facebook news feed or timeline to show that they physically checked in to the location. This could apply to things like hotels or restaurants.

Initiating a checkout would occur when a landing page is clicked.

Payment is based on completed registrations to a given event.

Actions can include engagement, clicks, or conversion.

Lead events are tracked by the pixel on a website and attributed to the ads.

The number of times people visit the actual business and can be attributed to the ad.

The number of search events can be tracked on an advertiser’s website.

A mention happens when a person types @ followed by a page name, then selects a page name from a list, in a Facebook post or comment.

Photo views are counted only when the image is clicked on.

A wishlist is essentially a shopping cart where the ad buyer pays based on adding items to the list.

The cost is based on the number of people who responded “interested” or “going” to a Facebook event.


Advertisers on Facebook (and Instagram) can target potential customers by location, demographics, interests, behaviors, and connections (people who like an app and their friends). Targeting lets advertisers build custom audiences, which is much more difficult to do in most media platforms. They can also create “lookalike audiences” where the target could be similar people to one’s best customers.


The Ads Manager is the basic interface that lets users create and run ads, target, set a budget, track performance, and see billing information.


The Power Editor tool is meant for larger advertisers who create and control many ads. It looks essentially the same as the ad manager from a user perspective, but lets the user manage multiple campaigns at once.


Similar to Power Editor and Ads Manager, the platform is free to use. It is meant for companies that need more than one account, want to manage multiple payment methods in the same place, assign permissions, or work with an agency or other partner. The Business Manager includes the Power Editor and Ads Manager in the platform along with audience insights (detailed demographic tracking).


Martin Soler runs his own consulting firm that works with company founders and management to build reputation and brand awareness. According to Martin, what drives him is marketing hotels and travel tech startups using great graphic design, communicating clear selling points, understanding large amounts of analytics, deciphering buying funnels, and measuring ROI.

Skift: Can you walk us through your background?

Martin Soler: I come from the marketing and publishing world. Prior to that, I was a general manager for two boutique hotels in Paris and managed those. I also did a bunch of other jobs in the hotel, including the more boring ones such as night audit and front desk. After being GM, I joined a hotel marketing agency and started doing marketing for hotels, first as a consultant, then as VP of sales and marketing of the agency.

I personally worked with over one hundred hotels from Paris, New York, Argentina, the Pacific, and so on. I have been everywhere. More recently, I was the CMO of Snapshot.

Currently, what I do is mostly focused on getting startups and hotel companies onto the market. My focus is hotel tech mainly, but I always try and keep one hotel client going at all times so that I also see the other end. Hotel tech tends to become a bubble if you are not in touch with the client side.

Skift:   Hotels and online travel sites typically relied on Google to generate web traffic and television and print to bring brand awareness. It seems like Facebook is becoming a bit of a hybrid where in some cases it replaces TV and in others, it could take some ad dollars from Google. When you look at traditional ad formats on Facebook, and now the dynamic ads for travel, how do you see that impacting the the travel industry? What part of the booking process do you think is gaining traction, how effective do you think Facebook can be as a tool for hotels, and OTAs?

Martin Soler: Today, Facebook is still a long way from taking money from Google. It is as you say a bit of this odd hybrid model. Google is so close to the end of the funnel, so it is easy to track the last click, which is what most people use.

Facebook is more of an inspiration or discovery and brand tool. It hasn’t yet found that sweet spot where it generates a lot of money, but people are using it a bit now. I think the fact that has now gone on to use Facebook has given them credibility.

For the hotels, targeting is there, but it is complicated and you don’t have intent. That is the thing with Google — they know intent.

Skift: How effective has retargeting been so far?

Martin Soler: Facebook is great for retargeting, because of the hours that people spend on it. The problem with retargeting for most hotel groups is they are doing it on a very crude level compared to what the OTAs are doing.  The hotels don’t necessarily track the dates (and they should) meaning that they keep retargeting after the person has already booked their stay. Whereas when you look at the OTAs, they factor all of that into their ads.

Skift: Google’s CPC auction is simple in what it is, but very complicated to optimize. We see a lot of the smaller hotels hiring third parties to run the ad campaigns. For Facebook, do hotels typically try to manage the Facebook advertising platforms themselves or do they more often partner with specialists to do that?

Martin Soler: What I have seen so far is that we have a few that try and do it themselves, but very few. It has become as complex as Google to manage ads on Facebook. You can get really granular and the minute you need to get granular you need somebody to help you.

Skift: Will even the mega chains like Hilton and Marriott partner with vendors?

Martin Soler: Yes. The process is becoming increasingly complex.

Skift:   What advice would you have for the hotel industry on advertising best practices?

Martin Soler: The thing with the hotel industry is that it’s really difficult to market a pure brand unless you are one of the major luxury brands, which really have a brand standard. A Hilton Tel Aviv or Hilton New York versus a newly built Hilton in Copenhagen are completely different products.

It’s not like a product company that can advertise a single product, which is going to be the same no matter where you go. This is the issue that the industry has and they are kind of solving it by making all of these sub-brands, but they also need fresh and consistent service.

I am seeing brand campaigns work for some luxury brands like the Four Seasons or with the very funky, modern brands like Citizen M. In both cases, they really do have a brand standard.

Skift:   Moving back to Google, where does Facebook win out?

Martin Soler: Google needs improvement at inspiration. They have not solved the issue of getting people inspired enough to come there and I think that is what they are going to try and do with their Google travel products, but they are a long way from it. Facebook is much better there, but doesn’t have the conversion. The difficulty is when you are measuring ROI, which is today what most advertisers are focusing on. Google gets more of it than Facebook in terms of pure ROI, because they are down the funnel and Facebook is the beginning of the funnel and sometimes Facebook is six months ahead, six months up the funnel, which is way too long to measure any ROI.

Skift: What do you think when a company like Expedia says that they see Facebook having Google-like advertising spend?

Martin Soler: It is going to take a long time, they haven’t even started reeducating users to use Facebook for search and until then, making Google-like advertising spend is going to take a while. Search is the holy grail of marketing online, but if you follow their design blogs, they are iterating on their search product a lot to try and get it to work, but they yet have to get anybody to actively say “I am searching for a hotel on Facebook,” It is just a foreign concept.

Skift: Messaging seems like it will first probably be for customer service and then down the road could be more widely used for booking. Given how much Facebook spent on WhatsApp, it seems like it could be a big monetization push where they license the technology or there are advertisements on the messenger platform. What are your expectations?

Martin Soler: I think that Facebook Messenger, WhatsApp, Google, and iMessage are going to be the platform, and then there will be channels within the platform very similar to how WeChat did it in China. They would have their credit card on file and then it is a payment platform, it is an everything platform. That is most likely where it is going to go.

I don’t know how they will monetize it other than doing revenue share on everything that goes via them, once they have got credit card data and people paying through it, they can take a transaction fee and that is it.
Frederic is a digital marketing and strategy expert focusing on the travel industry. His company, Gonzo Marketing, provides marketing expertise and counseling, from strategic planning (marketing plan, communications plan, etc.) to tactical planning — including promotional, advertising, and PR campaigns — as well as brand positioning. He also has been partnering with the SkiftEDU team on “How-To” columns on digital marketing for travel brands and conducting premium webinars.

Skift: Can you discuss your experiences, what part of the advertising process you focus on, whether it be on the technical or on the strategic branding side?

Frederic Gonzalo: For the past five years, I’ve been consulting, speaking, and training on everything that relates to digital marketing and almost exclusively focused on hospitality.

Usually what I do with brands is help them make sense of the digital landscape providing guidance on website audits, social media presence, strategy, email marketing campaigns, and website optimization. It’s mostly from a strategic standpoint and I’d say the only exception to that is perhaps with Facebook. I guess that’s for personal interest as well. I’m very active on it and interested to see the development they’re doing, not just for organic tools on how you promote, but also how the platform publishes content, images, and videos.

Facebook is probably the only tool that sometimes I will help clients understand the nuts and bolts of how it works and getting a campaign in place. I don’t necessarily do that with Google AdWords or Analytics and those are all tools that are available.

Skift: As Facebook rolls out dynamic ads for travel, do you see Facebook becoming a more important part down toward the bottom of the booking funnel or do you think it just becomes a stronger booking awareness tool that takes share from television and print?

Frederic Gonzalo: I think it’s mostly the latter. I think as much as Facebook is trying to position itself lower down the funnel, it boils down to how we use those tools. The day we’ll start going to Facebook as a reflex to search things, then sure we might be seeing a paradigm shift. We’re not seeing that, we still Google things, but we spend a lot less time on Google than Facebook.

I think time spent is where Facebook has a huge advantage, where we’re browsing, consuming content, and do it for forty minutes per day. You never spend twenty minutes on Google. You go on Google for a specific need. Maybe YouTube is more of a competitor as far as the Google environment to Facebook because you’re consuming content, you’re watching videos and you might go from one video to the next and then you consume content and you will see an ad.

From Facebook’s standpoint, I think it can be a useful tool for the whole remarketing aspect. You will still be marketing from Google, but I think we spend so much time on Facebook, that’s where the remarketing is. You were on the website for a hotel and then you got disturbed and you moved on to something else and then later that evening you’re on your Facebook, and then you’re seeing that travel package to Denver or to New York. That’s where it’s really powerful. Where you can drive conversion lower down the funnel to a more sophisticated tool such as e-marketing. Personalized or customized databases provide Facebook with some powerful tools.

Skift: Is Facebook more widely utilized by the hotels or OTAs?

Frederic Gonzalo: I think Facebook is good for everyone who travels. I’ve seen restaurants that are using it to drive traffic for lunch time or they’re promoting a new recipe or contest. They’re driving with the geo-targeted objective to get people to clear their itinerary, to find out how to get there. One thing that I’m seeing especially for events and attractions is where they’ll do coupons for example, the third person gets 50 percent off.

For destinations, I don’t see this much. Destinations seem to be using it in a very traditional way, with reallocated budgets to social media.

Facebook ads democratize things, where you can be a small page with 300 fans versus an airline that’s got 9,000 fans and you’re on the same playing field. You basically have the same tools to lead that war to get the coins. In that sense, I’m seeing pretty off-court OTAs and those guys are putting the big bucks, they’re transferring money that used to be traditionally put into Google Ads and seeing good results on Facebook.

Skift: On Google CPC seems conceptually straightforward, but optimizing it is complex. This leads to many advertisers outsourcing ad campaigns to ad technology vendors that specialize in that. On Facebook does that tend to be the case as well?

Frederic Gonzalo: I think we’re at a crossroad now. I think we’re certainly, and we are now and have been for the past year or so, because it’s so user friendly that we’re seeing a lot of brands especially the small inn mom-and-pops or the hotels that say, “I can do this myself.” It should to some extent, however; it has become so sophisticated.

It has become perhaps even more complex than even AdWords. If you want to do A/B testing, it’s a little quirky if you’re using AdMeter and you want to go into PowerEditor. Not everyone is necessarily comfortable using PowerEditor. It was maybe two or three years ago, that pretty much anyone could get in and build the ad.

Now, ad campaigns require more sophistication and more and more are using a third-party agency, community manager, or an internal expert. There are so many variables that come into play that indeed I think we’re going to see that shift in 2017 where for basic stuff yeah you can do it yourself but as soon as you want to etch your more sophisticated goals or audiences, then you’ll want to go to an expert.

Skift: Turning to Messenger and WhatsApp, how important do you see messaging becoming in the hospitality industry, both for the hotels and for the booking sites? Can that become a key way that Facebook finally monetizes those platforms directly?

Frederic Gonzalo: Yeah. I recently wrote an article on seven trends that I see for e-tourism in 2017 and that was one of them. I actually think it will be one of the big trends up there. That and the whole live streaming app, like Facebook Live. For sure, the messaging and the customer experience tool we’re already seeing it, we’re already seeing a whole bot and how he automates some of his processes coming into play with the OTAs.

Millennials are becoming one of the most important demographics for hotels in the U.S., so you have to be there with the tools that the millennials will want to use. There are quite a few 40- and 50-somethings also active here.

If you’re seeing more and more people going toward messenger, well, guess what, the reflex is there, they have one billion active users, so people have to read this thing. Well, sure I’m going to contact my car rental company or my hotel or the OTA, whoever I booked with and see, can I change my reservation, is there a restaurant at your hotel, so on and so forth. You will expect an answer.

Skift:   How do you see Facebook’s position in the travel industry playing out? Looking into the next decade, do people ever actually use Facebook as a travel search engine or is it the case where it’s brand awareness and it’s a nice consumer touchpoint?

Frederic Gonzalo: I can honestly say Facebook is trying to be the end-all, catch-all place where we spend all or most of our time. You’ve got to hand it to them, they’ve done a stellar job at staying on top of things and when they weren’t able to do things themselves, they would do it by acquisition.

Purchasing Instagram in 2012 seemed like a stroke of genius, getting WhatsApp, we tend to forget but they also bought Oculus Rift; that was a whole virtual reality thing that was pointed to become huge. Well, they just bought the number one player in the world. You’re seeing that and in the meantime, you’re seeing Twitter sort of dwindling and wondering where they’re going and they’re not having as much advertisement as they wanted and they had to shut down Vine. Periscope, which was Twitter’s, died because of the Facebook Live feed. With Instagram, it’s stories and Facebook more than Instagram, it’s going to start doing some Live stuff. Perhaps that’s going to kill Snapchat, but it is tough to tell because Snapchat’s got such a strong user base.

They’re really playing all the cards to have the ecosystem. For a travel brand, you basically don’t have much of a choice. Facebook is becoming the place where you need to be. Now, after that, I’m not sure about the whole e-commerce slide. It has always been there and they’ve been making efforts and now they’ve done catalogs of visual and the whole dynamic. We’ll see how that plays out. They’re not sitting on their laurels; they’re really doing things for our industry in advertising.

I think lower down the funnel you’re still going to need the traditional tools. Google Ads will always have an edge because people are searching for things. I think Facebook is basically covering everything else in between. It’s where you spend time, where you get the brand awareness, all the way to the conversion where there are specific offers for people where they may say, “this is cool, 50 percent off, I’ll click on ‘book now.’”

Skift: What about video?

Frederic Gonzalo: Facebook Live is really adding features and has got steam going this year and we’re going to start seeing more and more brands using it and putting subtitles in there. They will likely be using it to launch products and services. Instagram is supposed to go live as well. Once that happens, Twitter might have a hard time. I think Facebook and Instagram will own live streaming. I see potential for the brands to come up with advertising possibilities around this. It’s not there right now, but I can pretty much foresee that in 2017 they’ll have some sort of advertising around promoting your live streaming of whatever it is that you’re doing.


ReFUEL4 is the world’s first crowdsourcing platform for ad development, built on the Facebook and Instagram APIs. Using a pay-for-performance model, ReFUEL4’s base of more than 10,000 designers can create static and video ads in a matter of hours to sustain campaign performance. Using compositional and data analysis, ReFUEL4 provides partners and designers live campaign data that guides the further development of ads. This means that the ads improve as the campaign progresses. Skift conducted a lengthy phone interview with two of the co-founders in Asia. We provide key takeaways below.

We asked ReFUEL4 what they actually meant by: “Using convolutional neural networks and advanced algorithms, we process and benchmark performing creatives more accurately.” The team discussed how Facebook’s vast data lets ReFUEL4 marry that with the creative side. They scan each creative piece to see what is actually in it, and then compare that to data in Facebook and track the actions taken by viewers. The convolutional network looks for patterns in data and then makes predictions based on that. If data show that a picture of the Eiffel tower is effective at driving bookings from New York, they tell designers to include that in the creative side.

When asked about Facebook’s place in the travel industry and booking process, ReFUEL4 believes that Facebook will be strong in the whole booking funnel. Facebook has all the people and all the data and in a mobile-first world; ReFUEL4 believes Facebook will be the leading digital player in travel over time.

ReFUEL4 confirmed our interpretation of tracking where Facebook is able to track browsing habits even when a user is not logged into Facebook (on any site that partners with Facebook in some capacity) where the data is then used for retargeting. Facebook previously relied on Criteo and others for retargeting, but now is doing this mostly in-house.

When comparing hotels with OTAs, ReFUEL4 finds that OTAs are much further ahead on using Facebook. The hotels are very much in the experimental phase. The hotels need to start approaching Facebook as a real platform and using the retargeting data.

Measuring the path to purchase when you are not measuring the last click has been difficult to do. How to attribute this path and then spending dollars effectively is a challenge that advertisers are addressing.

Cross-device tracking has become seamless for Facebook across devices and time.

Conor has spent the last 12 years working on product innovation and startup companies. He led research, development, introduction of products to market, and client services for those products. These include products for Facebook, social media, mobile, radio, online advertising, music, general internet, and several areas of telecoms.

StitcherAds is the first SaaS platform specifically designed to help online retailers optimize their Facebook direct response advertising. The company “stitches” together a retailer’s first-party data such as product feed and site behavioral data and combines it with Facebook’s power at creating targeted reach to help the client put the right product ad in front of the right customer at the right time on the right device.

Skift: Can you walk us through your background and give a high-level view of StitcherAds?

Conor Ryan: I’ve been a software entrepreneur now for nearly 20 years. I used to lecture in a university, got into some research there and then kind of burst out into starting a commercial company, using some of that research. I’ve been doing that for years and my history is very strongly embedded in software, originally in telecoms and later the internet and now media.

We started the company as Betapond back in 2009. My co-founder, Declan, and I had just exited a previous company. At the time, Facebook was gaining momentum as the juggernaut it’s turned out to be. Back then, we focused on building the brand to engage customers and drive sales. As Facebook evolved, we evolved with them and took a very narrow focus. In late 2013, we just stopped everything else we were doing and built a platform that specialized in direct response on Facebook. We took a very deep specialization in the middle to lower end in a direct response offering while everyone else was going after brand dollars at the top.

Initially, we focused on e-commerce and it was built to be self-service, all our technology is built to be self-serve first. We started introducing travel-specific features and automation in early 2015 when we brought TripAdvisor and some other clients onto the platform.

Skift: For TripAdvisor, what type of relationship do you guys have with them?

Conor Ryan: We’ve always supplied software and support services to TripAdvisor. We got in there initially with Dynamic Product Ads, which was the first version of Facebook’s dynamic offerings that had just come out, and we’d already built our own version of that before as a kind of a hack on Facebook. Facebook was really eager to get their travel clients working on that and they knew we had the expertise, we were definitely number one, number two in the world with the expertise on dynamic ads on Facebook. We kind of made that product work for TripAdvisor through some customizations and some hacking around Facebook’s dynamic offering to make that work.

Trip was probably one of the first to actually run a dynamic ad for travel on Facebook, our travel ads on Facebook. Then, we were constantly hitting Facebook with requests for additional features that were travel-specific and that, along with just the huge pressure to build additional travel capabilities on the platform to literally try and get some of those Google dollars, was kind of what led Facebook to build out dynamic ads for travel.

We’ve worked very closely with Facebook and a lot of our clients to figure out what the different features should be to allow the travel client to get the level of efficiency that they need to make Facebook work for travel.

Skift: With dynamic ads for travel, do you think it could help move from Facebook advertising from brand awareness down to the more transactional side?

Conor Ryan: Yes, so I think the challenge that Facebook has is that it has to convince the travel industry, and it’s still in the early stages, to try and attract them away from the more traditional media, from television and Google.

Trip has had to prove that there is return and I think that’s probably why the first verticalized product we’ve seen had been at the retargeting level and the dynamic ad for travel, as opposed to anything at the top-level branding and I think that’s specific for branding control. For branding, it’s a much more generic offering, but for the bid that we made was that when it comes to travel, Facebook’s going to happen. Very clearly, deliver highly cost-effective advertising media for travel clients and the way to prove that initially is obviously retargeting.

Facebook has a very unique offering there in relation to its dominance on mobile. And when you roll in the identity basis for targeting, Facebook is the company that has the biggest opportunity to grab your attention again because of consumption of content.

Skift: On the re-targeting, is Facebook able to accurately capture cross-device transactions?

Conor Ryan: They absolutely are and that’s because it’s identity rather than cookie-based. In a cookie world, you’re talking about probabilistic: we think that this is the same person because we stick these different fingerprints on these different cookies that we merge together. But in Facebook’s world, it’s true and completely accurate.

How it technically works in simple terms is a user will be targeted an ad and Facebook knows that that Facebook user ID has seen that particular ad from that advertiser. Then, later on a desktop (where they are logged into Facebook in some browser in the background), the actual conversion happens. When they book that hotel room, the Facebook page will fire the ID of that user and the ID of the room, the product that they bought. And that goes back into Facebook and Facebook says, “Okay, so we just had this conversion on this particular client, has this user seen an ad from our client? Yes, they saw this ad two days ago, okay, let’s attribute that conversion to that particular ad.”

And it’s not only that particular conversion after your ad and for that particular prompt. It is done at product-level tracking, because of all of the data that’s contained in these feeds that the clients have to give to us to push into Facebook to give them that data.

Skift: Let’s say a user views a hotel ad and then winds up two days later booking that hotel but say through an OTA. Do the ad dollars just go to the OTA? How would that work?

Conor Ryan: It’s completely dependent on the client’s attribution model, their multi-touch attribution and how sophisticated they are. There are various levels of sophistication. OTAs are very sophisticated in how they measure. Some of the suppliers wouldn’t be as sophisticated but in a lot of cases, a particular supplier might know where that booking actually comes through and they might attribute 15 percent of the value of that booking to the Facebook ad to the OTA.

Skift: Interesting.

Conor Ryan: So again, that’s completely dependent on the client’s actual measurement and attribution model.

It can be a bit of a science and a lot of clients that have worked with us have Facebook to do incrementality tests where they don’t run any ads on Facebook, they see the actual impact that that has on booking, then they run an ad on Facebook, measure the impact there and say, “Actually, you know what, for the month we ran ads compared to the month where we didn’t run ads, we got 15 percent incrementality, so we’re going to take that 15 percent and use that for multi-touch attribution.”



As much as Google dominates search, Facebook dominates the social media landscape. As of September 2016, Facebook had 1.7 billion active users and both WhatsApp and Facebook Messenger had over one billion. The following three leaders — QQ, WeChat, and Qzone — are all owned by Chinese company Tencent. Facebook’s Instagram platform also has 500 million users. Facebook’s huge user base is a strong barrier to entry and leads to more ad dollars moving to Facebook.

Source: Statista, As of September 2016

In 2015, Facebook accounted for approximately 84 percent of social media industry revenue.


Desktop spending is expected to stagnate in the coming years while mobile growth slows gradually to the 10 percent level with the net being high-single-digit industry growth as we look into the next decade. This growth forecast by Statista sheds some light on why Facebook is pushing hard into new areas within advertising like travel and why other potential monetization streams may become more important (messaging currently does not contribute revenue directly). Facebook transformed itself into a mobile-first platform after its IPO with rapid revenue growth at over 50 percent per year over the past few years.

Source: Statista, As of September 2016


Facebook’s rapid growth has not been at the expense of Google, which despite having a massive $90 billion in revenue in 2016 grew 24 percent in constant currency. This means that digital advertising growth does not imply a war between Facebook and Google, but rather it is a case of the two leaders continuing to dominate the industry and take share of a growing pie. There will of course be increased competition between the two, but both should continue to grow nicely into the next decade.

Source: Statista, Company filings

Global digital advertising growth is more mature than the social media subset, but that is largely a function of search (Google) being more mature than Facebook. The total market is estimated to be at close to $200 billion with Google accounting for almost half of that and Facebook at close to 15 percent.

Source: Statista, As of September 2016

Source: Statista, As of September 2016


Mobile has become an increasingly important channel for the consumer. In Kleiner Perkins Caufield & Byers’ 2016 Internet Trends report by Mary Meeker, she shows that time spent on media in the U.S. in 2015 was as follows :

Advertising Dollars vs. Consumer Time Spent on Media Vertical

There is still a large gap in monetizing mobile from an advertising standpoint where the percentage of advertising spent on mobile is much less than the time consumers spend using it. This continues to improve with more money moving away from things like print to digital advertising. The secular shift should continue to benefit both Google and Facebook.

The fact that mobile is 2x under-indexed versus time spent is important as internet usage and smartphone growth become saturated. Kleiner Perkins Caufield & Byers estimates that digital advertising grew 20 percent in 2015 to $60 billion in the US. Of that pie, Google grew ad revenue 18 percent to $30 billion and Facebook grew 59 percent to roughly $9 billion while the entire rest of the industry grew 13 percent to approximately $22 billion, giving Google roughly 50 percent U.S. market share and Facebook 15 percent. Together, they accounted for around 75 percent of industry growth in 2015. While mobile usage itself has matured with the smartphone market, advertising is still in the middle of a transformation, and if dollars follow time spent, there is a $30 billion market opportunity just in the U.S. Should Facebook and Google maintain respective market share, there would be close to $5 billion on opportunity for Facebook to capture and $15 billion for Google.

We spoke with Rob Torres, Industry Director, Travel, at Google about this dynamic.

Skift: Broadly speaking, digital advertising spend is still significantly behind viewership levels.  Several studies point to mobile being just over 10 percent of ad spend versus 25 percent of viewership.  Do you see industry ad spend accelerating towards mobile?  TV seems indexed to time spent now while print is far over-indexed.  Where do you see the shift coming from?

Rob Torres:  Time spent with digital – driven by mobile – is now equal to time spent with TV, Print and Radio combined, so I definitely expect to see ad spend accelerating towards mobile.

I can’t say exactly what budget the shift will come from, but digital enables brands to understand a consumer’s intent and thus be a more effective spend.  Unlike traditional media, we don’t need to give every consumer the same message.  We can understand whether a consumer is looking to book a room, or if they are considering a new loyalty program, or if they are still getting ideas for vacations and connect them with a tailored brand message.

And now with cross device advances in measurement and data, brands are able to stay relevant with the same consumer where they are spending their time so I expect the spend in mobile to only accelerate faster.


ComScore found that Facebook by far has the most engagement by minutes for the average user and the broadest reach. This means that user growth may become saturated in the U.S., but the value proposition to advertisers continues to grow.

Source: comScore

Forrester Research found that Facebook is by far much more widely used than its closest competitors in the U.S.; Facebook owns the top three platforms. Additionally, users average eight sessions per day showing the desire to stay connected and is a reason why engagement has steadily climbed higher even in the mature U.S. market.

Source: Forrester Research


Below, we walk through our estimates of how much the largest online travel companies spend on digital advertising, how much of that goes to Google and Facebook, our estimate for industry spending, and the likely trajectory going forward.



Priceline has provided online advertising spend and total advertising spend in its historic filings. Online advertising has accounted for around 95 percent of digital spend on average over the past seven years. In 2015, online advertising was $2.8 billion.


Expedia does not provide a pure digital advertising metric, but we can estimate it. We assume that 90 percent of advertising is on digital. We take Priceline’s 2015 93 percent as a starting point and assume that Trivago flooding the market with television spending takes the television share a bit higher than Priceline’s budget. This takes us to $1.9 billion in digital ad spend for Expedia. We note that this attempts to strip out non-advertising costs embedded in the SG&A line item like employee costs, affiliate fees, etc.


TripAdvisor does not really use television advertising, so we assume 98 percent of ad spend is digital; the remainder includes everything else, including things like print. We estimate that the company spent $504 million on digital advertising in 2015.


For Ctrip, we make a similar percentage estimate that we did for Expedia, adding $643 million to the digital ad bucket. Ctrip would spend mainly in China, but we felt it was worth showing its contribution to global ad spend.

The top four online travel companies likely spent $6.2 billion on digital advertising in 2015 as that line item once again grew 35 percent. This should continue to grow at least 25 percent per year through 2017, placing 2016 at around $7.7 billion and 2017 at $9.6 billion.

Source:  Company filings, Skift Research Group Estimates


Given the vast number of private companies and limited data, we attempt to estimate an industry spending range based on how much market share the big four OTAs account for. We believe a likely range is 40 to 60 percent. For comparison, Google has close to 50 percent share of the global digital ad market and Facebook has 15 percent. It seems reasonable that the largest OTAs combine to have a similar market share/contribution of travel digital spending as Google and Facebook do in total digital ad spend received.

At the midpoint, we see the travel industry as having spent approximately $12.3 billion in 2015 on digital advertising, and this should grow more than 20 percent for 2016, which could place spending at over $15 billion. For a sanity check on our numbers, eMarketer estimates that U.S. travel companies spent $4.9 billion in 2015[1], placing U.S. spending share at 40 percent of the total. Within the total digital market, Statista estimates that 30 percent of the total comes from the U.S. Given the maturity of the U.S. online travel market, it makes sense to us that it would have a greater contribution to the total solely in the travel bucket.

Source: Skift Research Group Estimates


From our conversations with online travel companies, hotels, and digital experts both in this report and previously, we believe that the broader travel industry spends between 70 and 90 percent of digital advertising with Google. This would place travel-related ad spend between $8 and 11 billion or 12 to 15 percent of Google’s (technically Alphabet) revenue. Looking at 2016 numbers, Google’s travel revenue is likely around $12 billion while 2017 should hit $14 billion.

Source: Skift Research Group Estimates, company filings, numbers are for digital advertising only


Both publicly on earnings calls and from our industry contacts, we continue to hear how advertising budgets are adding Facebook as a true line item for the first time. We believe the starting travel digital ad market share is in the low single digits. However, we see this growing to at least as high as its total digital advertising share, which is over 10 percent. This will not be an instantaneous move, but we see the case that over the next three years, Facebook could control over 10 percent of travel ad spend, and potentially more. Off of 2015 numbers, this could be close to a 10 percent incremental revenue growth opportunity.

Source: Skift Research Group Estimates, company filings

Facebook’s 50 percent plus average ad revenue growth rate from 2013 to 2016 has been driven by growth in the user base, time spent by users, and growing the number of ads (ad load). On its past few earnings calls, management stated that ad load growth will not be a driver of growth as the load amount is now mature. This removes one of the three key drivers of revenue growth.

On one hand, this makes the number of ads that the travel industry will place on Facebook capped. On the other hand, this valuable real estate continues to grow in value. Travel companies that can generate high ROIs will be willing to pay more than smaller verticals. Broadly speaking, pricing for Facebook ads continues to increase as advertisers shift investment to digital. We expect travel companies to be part of the increasing cost per ad trend for Facebook.



Despite its large size, Facebook’s user base has continued to grow 15 to 20 percent per year, surpassing 1.2 billion daily active users in December 2016. For travel companies, this gives them a huge audience to advertise to where branding can be done at scale and targeting can be done at a much more granular, targeted level.


Despite its large size, Facebook’s engagement level (daily active users divided by monthly active users) has remained strong at 66 percent. What this means is that two out of every three of the 1.8 billion monthly users is using Facebook on a daily basis. For a travel company, it means that ads retargeting a consumer soon after a flight is booked or hotel site is searched is likely to be viewed quickly after the initial event. “Recentness” is crucial for better conversion in travel ads. Facebook’s high engagement level increases the likelihood of viewership and conversion, and that in turn increases prices travel companies will pay for ad space on Facebook.

Source: Company filings


The key U.S. and Canada market drives over 50 percent of Facebook’s revenue, despite accounting for just 15 percent of DAUs. The reason is straightforward with the average revenue per user (ARPU) being much higher in these markets at $19.81 in the latest quarter versus $5.98 for Europe, $2.07 for Asia-Pacific, and $1.41 for the rest of the world. The higher ARPU is a function of the North American economy, maturity of advertising in general, and the high engagement level at 78 percent. That engagement level is even more remarkable with the user base accounting for around 50 percent of the total North American population.

Across the travel industry ecosystem, the North American market remains incredibly important with the mega hotel brands being primarily U.S. based and online travel companies deriving a large part of revenue from the North American customers. Expedia and TripAdvisor generate nearly half of revenue from North America. Priceline technically does not generate much with all bookings done on counting as international, but we suspect the number of bookings coming from the U.S. into Netherlands based on could be in the 25 to 35 percent level.

The importance of the North American market for the travel industry lends itself to supporting ARPU continuing to rise for Facebook.


Europe is a key market, especially for the large online booking companies that spend billions on digital advertising. has dominated in Europe and Expedia continues to push more into the region. Smaller players are also becoming more aggressive in growing their European lodging footprint. Lodging in Europe is far more fragmented, so the online booking sites become even more valuable there as aggregators of best inventory. With 255 million DAUs, a 75 percent engagement rate, and only a $5.98 ARPU, the Facebook platform is a cost-effective way to push more European expansion.


Facebook has 396 million DAUs in Asia and another 388 million spread throughout the rest of the world. Engagement here is 59 and 64 percent respectively while ARPU is $1.27 and $0.94. Travel companies based in the U.S. and Europe can use Facebook to enter new markets where the cost to test is quite low. Local participants can reach the right audience as well in a cost-effective way.

Source: Company filings


Facebook has quickly gone from the experimental phase for many travel companies to a growing part of the advertising budget with its own line item. We see the importance of the platform for the travel industry continuing to grow. Below we walk through the unique advantages that Facebook offers clients.


Retargeting is perhaps the most important part of the advertising puzzle for the Facebook platform. If a person is searching for flights to London on April 3rd to attend our Skift Global Forum, but does not book anything, hotels can still advertise to that person when they open up the Facebook app or use the website version. The attractiveness here is three-fold for the hotel.

First, a hotel can target an ad at someone it knows is actually highly interested in visiting that destination on a specific date.

Second, the cost to run that ad will be lower than broader mass targeted ads (television, print).

Third, the conversion should be higher than other formats given the specificity and relevance of the audience.

More simplistically, hotels can also retarget people visiting their own website.


We spoke to a number of experts in digital marketing and will explain the nuances of cross-device tracking in what we hope is an intuitive way for those who are not in the proverbial weeds of digital marketing technology.

A typical path for a vacation booking would be as follows:

  • After the background search on Google about what destination to choose, a person will typically check flights.
  • In this example, we assume the search was done on a browser.
  • As long as Facebook log-in information is saved in the background and the travel site partners in any way with Facebook, Facebook can capture the search activity; a person does not have to physically log in on each search.
  • Most of us do not book the trip on the first stage of flight research. We’ll often gauge pricing of a few destinations and talk to our family or friends or significant others about the results. Days may pass before we go back and look to book a trip.
  • In the meantime, when the person opens up the Facebook app on their phone (how most people use Facebook now), hotels can show ads to that person for the destination and date searched for. The ads appear in the news feed and given the relevancy, they likely will at least be opened and looked at.
  • Facebook uses identification at the personal level (i.e. Jared Wein’s Facebook account rather than a cookie for my IP address that guesses it’s still me). This allows Facebook to more accurately capture actions across devices.
  • The person books a hotel to stay for two nights and attend Skift Global Forum in London.
  • The hotel would have paid based on the click or impression, but when calculating ROI and effectiveness of ad campaigns, it would then attribute the booking to the advertisement.

What if I search on a mobile app?

If a traveler conducts a search on Expedia’s app instead of the website, we believe Facebook can still capture this. Expedia, which is clearly increasing its Facebook advertising, can retarget to its users based on “events” where Expedia will use Facebook to target people that clicked a certain button.

According to Sojern’s Mat Harris, “They [Expedia in our illustration] would add a Facebook app event to that button, so that every time a user taps that button, the app would share with Facebook certain details, including the device’s advertising ID. Facebook can create an audience segment based on those button taps, and the app developer can retarget those users on Facebook.”

What is not captured?

If a traveler does not have a log-in saved for the Facebook browser and does not have the app downloaded with log-in saved, Facebook cannot capture actions and retarget. With most people saving log-in data on their devices and apps, we believe the vast majority of users can be targeted.

Will the process be the same as airlines and online travel companies using Dynamic Travel Ads?

Given that several of the OTAs/metasearch companies already use Facebook by manually adjusting dynamic ads, we see the transition as largely seamless for the big players with some likely added functionality and ease of use improvements. For smaller online travel companies, having ads tailored for them will increase usage. On the airline side, the process should be very comparable to hotels. Whether or not ads can drive bookings for airlines however remains to be seen as the booking process remains largely price- and time-driven whereas lodging is more differentiated.


Travel by its very nature tends to be aspirational. Vacationers spend a great deal of time thinking about where they hope to visit, plan what they will do when they arrive, how they will get there, and where they will stay. Most travelers search by price, availability, location, stars, and reviews as a first cut. After a list is formed, hotels that come across as the most visually appealing tend to win out on the battle for clicks and ultimately bookings. Facebook and Instagram are visual platforms where hotels can differentiate themselves as they would on a more expensive medium, like television.

The hospitality industry needs to create ads with a mobile- and social media-first approach rather than simply repackage television ads. Ads should be vivid and catch a viewer’s attention quickly. For video, it should be effective without sound and over a quick time period.


Facebook allows advertisers to target smaller subsets of the population. All things equal, Facebook will charge less per ad for less users being targeted. From Facebook’s perspective, the sum value of smaller targeted ads will likely be higher than broader ad campaigns as the smaller the targeted population, the higher the conversion should be, and the more value an advertiser will place on a cost-per-action basis (CPC, CPM, etc.).


In smaller and emerging travel markets with less mature online travel industries, cost per action tends to be much lower as conversion and viewership will be low (as will the price of the booking for the hotel). Hotels and online travel companies can test new markets with various ads to see what works without spending much money. In fact, one of our sources in the online travel space informed us that some low-cost markets tend to be extremely profitable on an ROI basis, despite being small in absolute terms.


With most users utilizing Facebook and Instagram via mobile apps, the ability to show advertisements organically is important. The first reason is that ads have a much less intrusive feel as one scrolls through a feed on Facebook. The second is that if ad blockers truly become effective, in-app advertisements that would not be subject to blocking will become even more valuable.


As mentioned in a previous section, the time spent on Facebook dwarves other media platforms (social and otherwise). Advertising dollars will flow to where the audience is as the platform matures and ad creation and optimization becomes more effective.



While the attraction of Facebook’s new travel ads is clear, companies on the traditional lodging and transportation side tend to be slow to change versus technology companies like the big OTAs. We expect gradual experimentation with Facebook and Instagram at first. Once companies see success in both generating bookings and building brand awareness, then ad spend should accelerate. For the OTAs, we see more spending on Facebook, but a massive migration from Google is extremely premature. Pulling too much ad spend from Google could easily lead to share shifts among the large OTAs.


If a person is retargeted for a hotel ad after searching for flights, it is not likely that they immediately click a link and book a trip. Instead, days or weeks go by before a decision is made. If the person then goes and books that hotel on a computer without a logged-in Facebook account (at work for example) or uses a metasearch site or OTA, figuring out how much of the attribution goes to Facebook versus the meta/OTA versus it becoming a direct booking for the hotel is complicated. Facebook is able to do multi-touch attribution for clients, but clients wanting to monitor themselves will find this more challenging than last-click methods.


Google may lose some incremental share, but a massive share of ad spend is not going to be moving to Facebook or anyone else from Google. This places a limitation on how much Facebook can grow its travel business. Fortunately, coming from such a low base, moderate share gains from Google plus more from television, print, and other digital channels along with growth in digital advertising broadly will lead to strong growth in the coming years.


Similar to Google, the theory behind ad campaigns is simple, but optimizing spending is complicated. For Facebook, it is perhaps even more so since there is the quantitative optimization part, but also the creative side as would be the case for television. Some companies go at it by themselves while others use third-party vendors. In either case, figuring out the best way to use Facebook to build brands and convert views into bookings will only become more complex. Small hotels may run a few ads and get frustrated that “nothing happened” while large brands may set themselves up for failure if they get overly optimistic on conversions on a new platform.

Those in the industry who are patient in building ad campaigns on Facebook without focusing on immediate conversion numbers will be well positioned to take share in the travel industry.


Christine Warner serves as the travel head of industry at Facebook. In this role, she leads operational excellence, revenue growth, and product strategy to help travel marketers capitalize on the shift to mobile. With over one billion people on Facebook and Instagram every day, Christine’s team partners with the world’s largest travel brands to empower connections that drive business results every step of the journey across every device.

Prior to joining Facebook, Christine was the west coast director of a Viacom digital marketing team. She also developed advertising partnerships for Viacom in New York for five years. Christine launched her career in finance at NBC Universal and holds a bachelor’s degree in business administration from the University of Michigan. When not planning her next trip, you can find Christine exploring the Bay Area with her husband.

Skift: We have been speaking with online travel companies, hospitality companies, and digital marketing experts, and the case continues to be made that more ad dollars will flow to Facebook and Dynamic Ads for Travel should help accelerate the growth. Can you discuss how travel has become more important to Facebook in terms of ad revenue and how Dynamic Ads for Travel are being rolled out first with the hotels and then more broadly to the rest of the industry?

Christine Warner: We we’re incredibly excited to have launched Dynamic Ads for Travel last year. This was the first industry-specific ad solution built by Facebook. Several travel advertisers had worked with Facebook to adapt Dynamic Ads to their use case, so we saw a clear opportunity to extend this product into the vertical to address the specific dimensions that are important to travel advertisers: destination and time.

Dynamic Ads for Travel allows travel advertisers to upload their hotel catalogs to Facebook to enable retargeting and cross-selling. While the initial rollout was specific to hotels, we will be expanding beyond this category in the coming months. For example, with the addition of the destination catalog to the product in September, airlines can now also tap into Dynamic Ads for Travel.

Skift: Facebook used to be viewed as a Social Media tool that was great for brand awareness and driving aspirational views. With the new ad formats, technology improvements, and wider acceptance by suppliers, Facebook appears to be moving down the proverbial booking funnel to a more transactional platform company. How do you view Facebook’s role in the travel ecosystem?

Christine Warner: First off, we agree that Facebook is no longer just a social media company. We are a content discovery platform and a mobile platform where 1.8 billion people come to discover and interact with the things that matter most to them, including brands.

We recently commissioned a study, The Mobile Compass, to better understand travelers’ behaviors in a mobile world and we learned that from the moment someone is inspired to take a trip to the moment they book the trip, they are spending 43 days and going to 56 different travel related touchpoints across devices.

During the average trip planning period we tracked, travelers spent five times more time on Facebook than on travel-related apps, sites, and searches. Facebook and Instagram are with travelers every step of the journey across every device and hoteliers can reach travelers while they are planning their trips through engaging personalized advertising and also create memorable experiences for people to share.

Facebook’s suite of advertising solutions address the travel funnel from inspiration and discovery to booking. While travel marketers are investing throughout the funnel, we see a focus at the bottom of the funnel to drive guest retention, bookings, and loyalty acquisition. We are also working with our travel partners to tap into the power of Facebook & Instagram as discovery platforms, driving brand messaging at the top of the funnel and engaging the customer throughout their entire purchase journey.

For example, Westin Hotels used a sequenced video and carousel ad campaign to encourage travelers to consider it for leisure weekend getaways – resulting in a 15-point increase in message association and a nine-point increase in intent to book.

Skift: When advertisers are retargeting, what issues do they face when using multiple devices in the search process? If a person views a hotel ad on Facebook, but then books on an OTA a few days later, but books the hotel they saw on an OTA rather than the hotel website, is the attribution split? How is the share determined?

Christine Warner: First let’s level set on a key differentiating feature for Facebook. Our platform is based off of identity which means we have the power to engage with people, not cookies. Being a people-based platform allows Facebook to unlock the ability to market and measure across devices throughout the travel planning process.

For example, with Dynamic Ads for Travel, hoteliers can leverage their first-party cross-device intent signals to automatically promote relevant hotels to people on Facebook from their inventory, to retarget or cross-sell.

To illustrate this, you can consider if a person searched on a client’s app for a hotel in Miami and then did further research on the client’s site, that client could retarget that individual to serve them an ad across devices to complete the booking. We can also measure what device the ad was served on as well as where the booking was made; either the client’s site or app.

As mentioned earlier, travelers are hopping across devices and travel touch points from the moment they are inspired to the moment they book. Most travel partners are relying on last click attribution but that doesn’t address the full picture. To address cross-channel, cross-device attribution, we are seeing partners move away from last click toward multi-touch attribution models to assign appropriate value across the channels that play a role in driving transactions.

This realization is why we regularly talk to our clients about the importance of putting the right infrastructure in place to set their business up to succeed — including multi-touch attribution models.

Skift: We have heard some nice things from your clients on how Facebook has been a very cost-effective way to advertise. One of the challenges on Facebook that we have heard from some is that it is hard to measure transaction attribution on their own without relying on Facebook to tell them how the ad performed. What are your thoughts here?

Christine Warner: Our number one priority is helping our clients drive the objectives that matter to their business. To do this, we have to have the right measurement solutions in place that our partners can trust. That is why transparency is critical to our business. We take this very seriously and have implemented a number of third-party integrations, which we’ve heard from our clients are very important to them.

Skift: Instagram has become an increasingly important part of the story. How is advertising there different from Facebook in terms of who the users are and how clients approach the platform?

Christine Warner: Facebook is more of a discovery platform in travel — an opportunity to seek guidance from friends, families, and brands on the right destination, the best hotel, the ideal itinerary. Instagram is about inspiration. The visual nature of the platform makes it an ideal canvas for travel — with the ability to convey the beauty and essence of a particular destination.

There are 1.8 billion people who use Facebook every month and 600 million people on Instagram. There is overlap in that people use both platforms. The biggest differentiator, again, is the frame of mind people are typically in when they approach each platform.

From the advertiser perspective, we’ve been excited to see travel marketers embrace both platforms to reach people where they are spending their time in a personally relevant way.

For example, Caesars Entertainment, the hotel, casino, and entertainment company wanted a cost-effective way to increase online bookings across its nine Caesars Entertainment properties in Las Vegas. The brand joined forces with Instagram partner CitizenNet to launch its Vegas Vacation Sale campaign on Instagram and in Facebook’s desktop and mobile news feed.

Because Instagram showed the highest returns, the company used CitizenNet’s programmatic budget optimization to reallocate its budget accordingly.

Using CitizenNet’s audience modeling, Caesars was able to identify the audience most likely to book a room. The advertising data agency used affinity interest targeting based on each of the nine Las Vegas properties the campaign promoted, and created lookalike audiences based on people who already booked a room online.

Caesars ran carousel ads that were tailored to each property and showcased the unique features of each resort. The ad for Caesars Palace, for instance, featured fun drinks in front of a tropical background. Ad copy encouraged people to “bask in the glory of Vegas vacation rates” and pointed to a “book now” call-to-action button.

Finally, the company used the Facebook pixel to determine the number of people who booked a room from its ads, as well as the total cost of each booking, which allowed it to measure the campaign’s success.

The campaign saw a 12.9x ROAS, was eight times more cost-efficient in cost per conversion compared to other channels, and saw double the return on investment compared to its goal.

Skift: The large OTAs have discussed how messaging and bots will become increasingly important. Some view this as an evolution and others a revolution. Given Facebook owns the largest platforms in the world, it would be helpful to hear what your thoughts are here. Will messaging be more about customer service or booking? Can you envision a world where Facebook follows a WeChat model and monetizes these platforms directly?

Christine Warner: Consumers have shown us their appetite to communicate on messaging apps — more than one billion people use Facebook Messenger. In travel, and in other industries, messaging apps are the new frontier of customer service. They offer an opportunity to engage one-on-one with customers to provide deeper connections and better, more personalized service.

We’ve seen travel brands jump into the messaging experience, experimenting with new ways for people to interact with brands within the app. For example, Expedia launched a booking bot and IHG has a messenger bot that people are invited to interact with after booking a stay. We expect to see more experimentation with bots and customer service, but we don’t speculate on future scenarios for our platforms.

Skift: How is Oculus being used in the travel space? We’ve seen it at conferences and airports where destinations can be viewed to create interest.

Christine Warner: We believe VR will be the next big computing platform. But it’s still incredibly early days here and we are focused on the consumer experience when it comes to Oculus.

Right now, we are guiding our partners to think about what they can do with 360 video, already available on our platform and a first step into the immersive world of VR. Delta produced a 360 video of an A330 taking off in the Mojave Desert to bring consumers inside an experience unlike anything they’d shared before.

Skift: The broader travel industry tends to be one of the slower industries to adapt to changing technology. Are you seeing ad dollars flow to Facebook from areas like print and television to the extent you have in other industries?

Christine Warner: We’re regularly seeing brands across every industry look to Facebook to complement their TV strategies. With 1.8 billion people on the platform, Facebook is helping marketers reach people who are now harder to reach on the big screen. In fact, we’ve seen that when TV and Facebook are combined, advertisers see a 19 percent increase in targeted reach versus TV alone.

Skift: If we look out into the next decade, can you see Facebook ever doing an instant booking type platform like Google is rolling out, essentially being a metasearch provider?

Christine Warner: We don’t answer hypothetical questions. Facebook is focused on building the best experience for people and value for businesses. For travel, this means helping travel advertisers connect with travelers who care most about their offering.

Skift: Turning to the bigger picture in travel, Facebook and Instagram are widely used to influence travel demand. How have some of the geopolitical issues impacted the broader travel industry and what, if any, is Facebook’s role in encouraging travel despite macro uncertainty in the world?

Christine Warner: We continue to be a place where people come to discover, share, and find inspiration around their travels.

Neasa leads the EMEA travel vertical, where she co-ordinates the EMEA travel team to develop the travel growth strategy on Facebook.  Her role is to work closely with travel clients to help develop their marketing strategy, offer best practice and drive their business needs through Facebook.

Skift:  Outside of the U.S., the hotel industry is far more fragmented with more independent and boutique hotels dominating the space.  To us, these types of hotels are a perfect fit for Facebook’s new Dynamic Travel Ads where they can target and retarget small subsets of the population.  How do you view the market for digital ads in the hotel space for EMEA?

Neesa Costin Bannon:  To date we have seen a lot of the global hotel groups have great success using this product.  IHG ( was one such example (the global hotel company switched to Facebook’s specialized dynamic ads for travel from regular dynamic ads, resulting in a 20 percent lower cost per booking and a 50 percent increase in scale).

Our DAT solution is a great product for any travel advertiser who has been investing in remarketing.  Dynamic ads for Travel allows travel advertisers to upload their hotel catalogues to Facebook to enable retargeting and cross-selling whether you are a boutique hotel or a larger hotel group.

Skift:  The broader travel industry tends to be one of the slower industries to adapt to changing technology.  Are you seeing ad dollars flow to Facebook from areas like print and TV to the extent you have in other industries yet?  Are the dynamics different across regions?

Neesa Costin Bannon:  We’re regularly seeing brands across every industry look to Facebook to complement their TV strategies.  With 1.86 billion people on the platform, Facebook is helping marketers reach people who are now harder to reach on the big screen.  In fact, we’ve seen that when TV and Facebook are combined, advertisers see a 19% increase in targeted reach versus TV alone.

Skift:  In terms of the ad campaigns, are you seeing successful direct conversions in your markets in a meaningful way or are we earlier days with brand awareness being the main factor?

Neesa Costin Bannon:  Facebook’s suite of advertising solutions address the travel funnel from inspiration and discovery to booking.  While travel marketers are investing throughout the funnel, we see a focus at the bottom of the funnel to drive guest retention, bookings, and loyalty acquisition.  We are also working with our travel partners to tap into the power of Facebook & Instagram as discovery platforms, driving brand messaging at the top of the funnel and engaging the customer throughout their entire purchase journey.

Skift:  In Asia, more consumers tend to book travel via messaging platforms.  Do you expect Facebook Messenger and WhatsApp to play a bigger part of the booking process or will they be more about customer service where hotels and online travel companies use the platforms to communicate with guests and perspective customers?

Neesa Costin Bannon:  Consumers have shown us their appetite to communicate on messaging apps – more than 1 billion people use Facebook Messenger.  In travel, and in other industries, messaging apps are the new frontier of customer service.  They offer an opportunity to engage one-on-one with customers to provide deeper connections and better, more personalized service.

We’ve seen travel brands jump in to the messaging experience, experimenting with new ways for people to interact with brands within the app.  For example, Skyscanner launched a search bot and KLM were launch partners for the messenger platform where its used for booking as well as customer service.  We expect to see more experimentation with bots and customer service, but we don’t speculate on future scenarios for our platforms.

Skift:  What are some of the unifying attributes of the most successful travel ads run on Facebook?

Neesa Costin Bannon:  We can advise clients on best practice around targeting, tracking, placement, ad format and creative.  A successful travel ad on Facebook is tied back to the key business objectives of the client.

Skift:  The DTAs are first for hotels, but many online travel companies we spoke to like the ads so much they are just using them with workarounds.  When can we expect to see DTAs specifically for other sectors of hospitality and OTAs/metasearch sites?

Neesa Costin Bannon:  We are incredibly excited to have launched Dynamic Ads for Travel in May 2016.  This was the first industry-specific ad solution built by Facebook.  Several travel advertisers (Hostelworld and KLM for example) had worked with Facebook to adapt Dynamic Ads to their use case.  We saw a clear opportunity to extend this product into the vertical to address the specific dimensions that are important to travel advertisers: destination and time.

Dynamic ads for Travel allows travel advertisers to upload their catalogues to Facebook to enable retargeting and cross-selling.  While the initial roll out was specific to hotels, in EMEA we saw a clear demand for a flight solution due to the number of airlines in EMEA region and an airline solution was launched in December 2016.

Skift:  One of the unique benefits of Facebook for advertisers is the dynamic retargeting functionality based on ID rather than cookies.  Our conversations have been with developed markets industry experts where cross device tracking is largely seamless.  Are there different issues in emerging markets?

Neesa Costin Bannon:  Our platform is based off identity which means we have the power to engage with people, not cookies.  Being a people-based platform allows Facebook to unlock the ability to market and measure across devices throughout the travel planning process.

For example, with dynamic ads for travel, hoteliers can leverage their first-party cross-device intent signals to automatically promote relevant hotels to people on Facebook from their inventory, to retarget or cross-sell.

To illustrate this, you can consider if a person searched on a client’s app for a hotel in Paris and then did further research on the client’s site, that client could retarget that individual to serve them an ad across devices to complete the booking.  We can also measure what device the ad was served on as well as where the booking was made; either the client’s site or app.

Travelers are hopping across devices and travel touch points from the moment they are inspired to the moment they book.  Most travel partners are relying on last click attribution but that doesn’t address the full picture.  To address cross channel, cross device attribution, we are seeing partners move away from last click towards multi-touch attribution models to assign appropriate value across the channels that play a role in driving transactions.

This realization is why we regularly talk to our clients about the importance of putting the right infrastructure in place to set their business up to succeed — including multi-touch attribution models.



We believe that Google will continue to dominate the travel digital ad landscape despite the rise of Facebook. If we assume Google generated close to $11 billion from travel ads in 2015 and Facebook takes its market share up to 10 percent from three percent, there would be just under $1 billion in ad dollars that could move from Google if all those dollars are purely swapped out. While that sounds like a lot, we need to put this in context.

First, there is a chance that none of this even comes from Google and comes from other players in digital or television or print.

Second, $1 billion would only just exceed one percent revenue headwind to Google, given its massive size.

Third, the travel digital pie should continue to grow at well over 20 percent per year, adding $2 to 4 billion each year to total travel digital ad spend. While travel-related growth would slow, it would be about maturing rather than shrinking in absolute terms.

Fourth, the new instant booking/metasearch model from Google can, and likely will, replace some if not all of any ad share lost to Facebook.

The risk remains for Google that ad budgets migrate en masse to Facebook. However, if some companies do that, the ones that remain aggressive on Google would likely take booking share as Facebook is not yet mature enough to compete with Google on traffic acquisition. Again, we see a gradual share shift to Facebook where the large travel companies utilize both channels in different ways.


Online travel companies have come to rely heavily on Google to drive traffic. This makes sense as each company wants to be the one a user clicks when searching for something like “hotels in NYC in April.”

A quick Google search returns Kayak (Priceline’s meta site), (Expedia’s hotel OTA brand), (Priceline’s OTA behemoth), and then Trivago (Expedia’s hotel meta site). Unsurprisingly, there are not any hotels in the search results as Expedia and Priceline dominate bidding on AdWords. Moving down the screen, we see Google’s metasearch engine (Hotel Ads) returning three hotels. Clicking on the hotels actually then takes us to more Expedia and Priceline sites to book. The large hotel chains unsurprisingly compete more once clicking through the hotel ad (i.e. clicking a Hilton property would have a booking link along with OTA links, while small chains tend to not be one of the top link choices).

We searched in Rome as well and the findings were the same where every ad was from one of the big online travel companies. The largest online travel companies dominate Google search for lodging. The reason is twofold.

First, and most simplistically, Priceline and Expedia have massive advertising budgets and plow much of this into Google.

Second, and more subtly, when a consumer is searching for “hotels in…” they are by nature of the search brand agnostic. A hotel company trying to capture that search result and convert it will have difficulty as that traveler is looking to compare many hotels based on location, price, availability, rating, etc. A Hilton or Marriott is better off spending their budgets on unique characteristic searches like boutique luxury hotel along with spending on metasearch/OTA sites. Additionally, hotels should spend across channels to build brand awareness and loyalty, differentiating rather than trying to compete in AdWords on broad search queries.

For air travel, things are a bit different where the first ad is often an airline. The difference here is that the airline industry is far more consolidated than the hotel industry. When I search for flights to London, British Airways is the first ad. This makes sense, as I very well may just click that link if I am not going down the metasearch model to start. The next links are Kayak and specialized sites like cheapfareguru. Google Flights comes lower in the page with various flight offerings to choose from.


We see the online travel companies continuing to use Google to capture share. Facebook’s dynamic retargeting capabilities for hotels are already being deployed by the online travel companies despite the product not officially being for them (yet). We envision a consumer searching for flights to London on Kayak and whether or not they book that trip, Kayak (or retargeting that consumer with ads on Facebook. The benefit over a single hotel is that if the online travel company gets the audience to click back to the site from Facebook, they have the full breadth of its listings to drive conversion. For a hotel, it may be a specific property or a chain where even the largest chains have at most one to two percent of the OTAs inventory. If we add alternative lodgings, it falls to well below one percent.


Cheapflights Media acquired Momondo in 2011 and became known as Momondo Group; on February 7, 2017, Priceline acquired Momondo Group for $550 million to place it under Kayak[2]. It has sites in 35 markets in 20 languages, employing 350 people with offices in London, Copenhagen, Boston, and Sydney. The company operates with a dual-brand strategy with Momondo (founded in 2006) focusing on inspiring travel primarily in non-English markets utilizing both flights, hotels, and car rentals, and Cheapflights focused more on the English-speaking flight metasearch market. For more on Momondo, please see our 2017 Outlook on Metasearch in Travel.

Skift: Can you provide an overview of how you view the Facebook platform?

Hugo Burge: For us, Facebook allows for lower cost media, more targeted audiences, cheaper creative production, and more relevant and personalized content for branding than traditional branding channels that we can track right through to visit. It allows us to be extremely nimble and hyper targeted, which is especially useful when we are entering a new market or testing a new campaign or product, as it permits low-risk, low cost, get-in-and-get-out flexibility.

Facebook is used in various ways for both “softer” brand building and “harder” commercial purposes, combining top-of-funnel travel to our sites via lively, engaging, and informative content with a key focus on effective branded content, as well as a combination of reach and frequency ad products together with remarketing ads and custom audiences to create full-funnel marketing campaigns that allow us to sequence and map direct response to drive customers down the purchase funnel from awareness right through to conversion.

How that blend works, though, varies according to circumstance, market, campaign objective, brand, etc. There’s no single formula for success, but the benefit of the platform is that it’s flexible enough to accommodate that need for a varied approach.

For us, Facebook is already important for diversifying traffic and direct response activities from search, as well as proving a really effective branding channel and we can see that as user behaviors evolve and the channel further matures, that Facebook will take more of the pie.

It helps that Facebook itself invests in product development specifically for the travel vertical. DTAs (dynamic travel ads) are a good example of Facebook creating specific ad products to help travel partners drive direct response revenue. Facebook has long offered strong branding benefits to challenge offline and other traditional channels, and we believe that this will only increase as brands get a better view on attribution and the impact Facebook has on the end conversion when customers engage with Facebook at various touchpoints further up the funnel. The data we have to do this is only going to get richer.

Skift: In terms of Google versus Facebook, from our conversations it seems that in the broader industry, 75 to 80 percent of digital ads are on Google? Is that a similar case for Momondo?

Hugo Burge: I would say that I think there aren’t rules as percentages, and I don’t think we’d be comfortable giving percentages. I’m aware of companies in travel that do no advertising on Google and just use Facebook. I’m aware of some companies who just do TV without other channels. I think everybody has their own strengths and weaknesses. I don’t think there’s a generic rule.

What I would say is that Momondo is incredibly privileged to operate in a global environment with Facebook and Google as global channels that can reach enormous audiences. I think Google is more proven as an ROI generator, but I think Facebook has been catching up fast in terms of offering ROI driving opportunities. Perhaps its unique selling point is our ability to project a brand in a cost-effective manner to create highly targeted and differentiating campaigns, in an environment that lends itself very well to travel and travel inspirations, and word of mouth, and generating a virality around travel decisions and brands. I think Facebook has carved itself out a unique niche, and I think that everybody would agree that Facebook is growing faster than Google. It’s an earlier stage of maturity.

I think everybody is learning. I’m certainly learning all the time, and I think our experiment and our experience are very much based on a test-and-learn basis within Facebook in particular, and our strategy is evolving quickly.

Skift: On the DNA journey, if you could just walk us through the creative process behind that — how that became so popular and how it was rolled out on the various digital channels.

Tore Pein Jensen: It was just a brilliant idea and concept that we came up with, together with a local creative agency and my team here. The whole idea was that we wanted to make something that was as strong as possible on the creative side. It turned out really well, so we shot the film and made the production here in Copenhagen, and then both the YouTube and Facebook guys said that a five-minute viral film is never going to fly. But we believed so much in it, we gave it a try and it went really well. It certainly outperformed everything that we thought.

We had divided the budget for trying to get this up and flying into YouTube or Facebook as the main video platforms. But there’s no doubt that Facebook, already after the first couple of days was much, much stronger than YouTube. Almost all the views that we have are via Facebook or the Facebook platforms in various ways. I would say it totally outperformed what we actually thought in terms of YouTube coverage. We did put a lot of effort and money into boosting it. Normally you would see companies put a lot of money into the first couple of days, really do it to get some traction and to get it spread out there. We saw a fast, natural, and organic expression from the video. We have today almost two hundred million views on the video, which is really awesome.

Skift: Of the two hundred million, how much of that was Facebook versus YouTube versus others?

Tore Pein Jensen: When you look at it, there’s actually three main drivers for the video views. YouTube is one driver, then there’s our platform where we have the media also. Then there’s Facebook. More than 80 percent of all the views came through the Facebook platform.

Skift:   Great. For a campaign like that where you’ve got a really great brand awareness, did you guys see an uptick in traffic or bookings after that campaign?

Tore Pein Jensen: The only purpose was actually to start conversation out there. I’m not so glad to talk about the actual performance numbers, but I’m always asked. It’s no secret that we saw an amazing uplift. The first couple of weeks we saw a growth in 48 percent of our total website traffic. That’s a lot for this global Momondo we’re talking. We saw a huge uplift in traffic, and a tremendous uplift from just Facebook, of course.

Skift: In terms of how you position Momondo via Facebook across the different markets, are certain markets more receptive than others? Is there any consideration there in terms of how you apply, deploy Facebook as a marketing strategy across the different markets?

Tore Pein Jensen: Yes, there are. There’s as many strategies as there are countries. We obviously find some markets to be more profitable than others.

Hugo Burge: We don’t want to give away too much about what’s working for us and what isn’t, but I think it’s a great point. Some markets are better value than others for us. So again, it’s testing them. A huge advantage of both Google and Facebook is that you can enter new markets in a relatively cost-efficient and proven way, and benchmark against other markets, to find out where you’re getting value. It continues this theme of online marketing offering phenomenal transparency, and phenomenal scalability when you find things that are working.

We’re incredibly privileged to work in an environment where there are global channels that allow us to grow internationally. This is the age of global business. It’s an extremely exciting environment to work in.

Skift: That’s really fascinating, just thinking about drilling down into the country level, and testing different markets or different regions and cities, and the cost effectiveness of those campaigns and using Facebook and Google as a way to understand what the opportunities are, even in smaller markets. It’s probably safe to say that a marketing team would take it to that level, right?

Hugo Burge: Right.

Tore Pein Jensen: We love data. You could be very nimble with Facebook. I think you have to get to a point where targeting can be so precise, versus the return on that investment in terms of how much value you’re going to get from such a small niche. It gives you that flexibility, and the ability to get in and get out quite quickly if you find it’s not working. You can change tactic and move on and try different creatives, a different copy. It has phenomenal test and learn flexibility.

Skift: When you guys are optimizing the platform on Facebook, figuring out which ads should be targeted when, is that all done in-house, or do you guys partner with third parties?

Tore Pein Jensen: It’s in-house. We decided many years ago that if we needed to beat the competition, then we needed to do things ourselves.

Skift:  What about Instagram? You’re a very visual brand. Are you using Instagram?

Tore Pein Jensen: Yes, we are.

Skift: How are you guys in terms of when you play your ads on Instagram? How different is the approach versus what you’re doing on Facebook?

Tore Pein Jensen: It’s really different. It’s much more inspirational. It’s much more competitions. It’s sharing about some of the more high-funnel activities that we’re doing.

Skift:   I think it’s been a great overview. Hats off to all the efforts that you’re making and putting into the platform. It’s a very unique experience in our opinion, in terms of what you guys are doing with Trip Finder, inspiration, and the overall experience online. It’s great to get your insights on this critical piece of the puzzle here in terms of Facebook and how a brand like yours is using the platform. Is there anything that you think that we didn’t discuss that’s important that you’d want to mention?

Hugo Burge:  Facebook should be seen as multi-channel in isolation. I think you’ve highlighted that the Momondo brand comes across as very inspiring. That’s very deliberate. Our primary company goal is to create product that users love. The Momondo brand is expressing not only that it’s a better travel search product, but it’s a more inspiring travel product. I think that what you’re tapping into and seeing is an extension of the brand being played through Facebook, which is a particularly good channel for expressing those values.


Recently acquired by Ctrip, metasearch site Skyscanner was founded in 2003 with a focus on airlines. The model has evolved over time to now include hotels and cars, but airline metasearch remains the overwhelming majority of the business. The website had over 50 million unique monthly visitors in 2015, which was up nicely from 30 million in 2014. The company has localized sites in 31 languages in 70 countries, employing over 800 employees in 10 countries. For more on Skyscanner, please see our 2017 Outlook on Metasearch in Travel research report.

Colin Kim looks after Skyscanner’s paid media investment channels, as part of an in-house team looking after display, search, social, and retargeting globally.

Skift: How does Skyscanner view Facebook for advertising for brand awareness versus a more transactional/booking driver like Google? Does Facebook drive bookings and traffic meaningfully yet? Or is it still early days?

Colin Kim: Over the past few years, we’ve seen a shift with Facebook ad products becoming versatile, addressing both brand awareness and performance driven objectives. As a result of this, we’ve used Facebook’s ubiquitous reach of users and innovation in ad products (i.e. dynamic product ad or dynamic travel ad) as a new user acquisition tactic as well as a tool to capture intent.

Skift: Following the Ctrip acquisition, how will that impact advertising on Facebook given censorship rules in China? Will Skyscanner be independent enough that it will not matter?

Colin Kim: Skyscanner is operationally independent to Ctrip, so we don’t see this impacting advertising globally.

Skift: How important is Facebook likely to become in the travel ecosystem? Could it help alleviate advertiser’s reliance on Google or is it more taking dollars from print, TV, and other digital channels along with the whole digital pie just growing?

Colin Kim: We certainly believe that travel is an inherently social activity, and as such, Facebook will continue to play an important role. In terms of reliance, our investment is focused on channels where performance and impact can be measured, and our digital investment has been growing at a rapid pace across various channels.

Skift: Facebook is expanding with Dynamic Ads for Travel. How effective do you see this becoming?

Colin Kim: We see this as very effective. Dynamic Ads for Travel is one of the first mainstream ad products that we feel was designed with the needs of travel brands in mind. In our opinion, its features are ahead of its time with many that are not available through traditional retargeting vendors. However, success will depend on ability to predict user behavior and recommendations for prospecting.

Skift: When you advertise on Facebook and Google, do you typically manage those campaigns in-house or use third party specialists?

Colin Kim: We manage those 100 percent in house. We feel strongly that aligning our internal teams and maintaining knowledge in-house is crucial for continued success.

Skift: On the retargeting side, does Facebook’s ID tracking make it seamless to track bookings across devices? Or is it still a challenge?

Colin Kim: While Facebook does have a unique advantage with deterministic cross-device identification capability, it’s difficult for us to rely solely on this data as our users may have multiple paid channel touchpoints.

Skift: What areas do you want to see improved as Facebook matures for the travel industry?

Colin Kim: It would be beneficial if brands could leverage user-level audience data management, enabling a greater integration of first-party data. We have a rich data set of user behavior across travel searches and bookings, yet this is difficult to use in targeting audiences on Facebook, above and beyond the basic capabilities (i.e. custom audience) offered.

Skift: Can you discuss how Skyscanner is able to target and retarget very specific audiences?

Colin Kim: We use a proprietary user segmentation tool that analyzes the state of awareness through purchasing intent. This segmentation helps fuel our decisionmaking for how we create audience cohorts for remarketing efforts, as well as look-alike audiences for the most valuable users.

Skift: When we look out into the next decade, how large in terms of digital share in travel do you see Facebook becoming?

Colin Kim: As Facebook continues to develop new ad offerings to meet the needs of travel advertisers, we see it taking a larger share.

Skift: In terms of ROI, how does Facebook rank versus other platforms?

Colin Kim: We continue to invest in platforms where we see positive contribution to the revenue bottom line. But of course, different tactics have different expectations and we keep those expectations in mind depending what our objectives are. For example, we would expect positive ROI in a very short time frame for retargeting, whereas with awareness-building activity, we look at acquiring a new user and their lifetime value over a longer horizon.

Skift: Have you found much different conversion and success rates in different countries? Where have you seen the most success in Facebook so far?

Colin Kim: Yes we have. As Facebook continues its focus on mobile users, we see higher success rates in markets where traveler booking behavior has shifted to mobile, in markets like Korea, Singapore, and Australia.


Much like the OTAs and metas rely so heavily on Google, the hotels do so on the OTAs and metas. Hotels also utilize Google to drive bookings by bidding on specific search terms and now on Hotel Ads, but it tends to be more about the OTAs and metasearch sites. There is little controversy that hotels can effectively increase bookings on third-party booking sites, but often times, hotels believe the costs are too high and are now running aggressive marketing campaigns to attempt to push direct bookings.

With dynamic travel ads initially being click- or impression-based rather than commission-based, boutique hotels especially will need to be careful about overextending Facebook ad campaigns. Hotel groups that are used to managing things like metasearch ad spend will be well positioned to jump into Facebook’s dynamic travel ads, where effective campaigns can be lower cost than lower risk cost per acquisition ones elsewhere. We say lower risk because there is no payment until a booking occurs versus CPC where payment is on clicks that may not ultimately convert.

We also see Facebook as a very strong brand awareness tool that can be combined with television campaigns. Messages should be similar, but delivery should be much different with each campaign on Facebook being designed specifically for mobile digital media.

We also expect that hotels will retarget consumers by utilizing different types of traffic to help generate bookings.

  • People who came to the page, but did not book

Here, if the hotel works with Facebook, it can track users to the site by a Facebook unique identification (so long as the user is logged into Facebook in the background). The hotel can then show ads to that user on Facebook that may present a special discount on a very relevant offering and see high conversion on the ad.

The hotel could also group people coming to visit the site and then segment that population to run different types of branding or aspirational awareness ads.

  • Flight and destination search

Because Facebook is tracking user behavior across many sites that partner with Facebook (including the OTAs), it has flight search and travel intent insights that hotels would not have on their own. A hotel can partner with Facebook to retarget consumers who were searching or booked flights to a given destination. These people are clearly more likely to be ready to book a hotel given a commitment to a travel destination and date. Showing ads to that subset of the population can be an effective way to drive bookings where the CPC, CPM payments can be effectively lower than what a hotel would pay the OTA.

IHG for example is using Dynamic Ads for Travel to highlight benefits of its loyalty program, IHG Rewards Club, and also increase bookings on its sites. “We are looking to engage customers based upon what they have researched on our IHG brand websites,” said Michael Menis, IHG’s senior vice president of global digital and voice. “We can focus by brand, location, and stay dates with the exact live pricing. We are also in the process of adding ways to highlight hotels that may best suit travelers’ preferences. Carousel ads are used to show multiple properties or markets in aggregate from searches within seven days.”

IHG said the new ads led to a 50 percent increase in scale, or the brand’s ability to reach relevant travelers who are actually looking to book an IHG brand hotel. It’s also seen a 20 percent lower cost per booking for the ads. “The ads have a direct effect in bookings generated as we’ve seen that we are able to place content in front of consumers toward the bottom of the purchasing funnel,” said Menis.[3]


Fiona leads all social media and digital marketing efforts for Two Roads’ brand including Destination Hotels & Resorts, Thompson Hotels, Joie de Vivre Hotels, Tommie Hotels, and Two Roads Hospitality with over 100 individual hotels under the Two Roads portfolio. She specializes in digital and social strategy planning and development, online community management, blogger outreach, content marketing, application building, fostering online and offline awareness for brands, on- and off-page SEO strategy, branding, social media strategy, establishment of strategic partnerships, and business development.

Skift: When you look at the Two Roads portfolio, how are the assets positioned in the marketplace?

Fiona Boyce: We have a really wide variety of brands. JDV hotels are fun and family-friendly. Thompson hotels are upscale, boutique luxury hotels. Destination Hotels is more of a resort brand. There are a few Destination Hotels that are smaller, but most of the Destination Hotels are larger ski, golf, or beach resorts. Then we have Alila hotels, which are primarily in Asia and those are very, very high-end, luxury resorts. Tommie is a brand that is launching in the next year or so and those are smaller, with the idea of a smaller room footprint with more programming and public spaces for people who really want to stay somewhere with a sense of community.

Skift: Can you walk us through how a typical ad campaign would work on Facebook?

Fiona Boyce: Let’s just say it’s a summer campaign. We’ll have two different kinds of ads. We’ll have the more traditional type of ad that will be retargeted ads and dynamics ads with the special offer and a book now button and those are running within the main Facebook feed. Then, what the social media and content team will do is promote the same details of the offer but rather than using a book-now button, a room image or something that feels like what you would see on a website, we will shoot our own content so it feels more editorial in nature.

We are supporting the same campaign, but we’re really using our social pieces to talk about it in more of a storytelling way, rather than just here’s a room and here’s a promo code and book now. Then we will boost those posts. We do a lot of influencer partnerships to also shoot a lot of content on-site at our hotels so that it’s not just our images, but it’s also coming from people who are actually experiencing what it means to stay at a Two Roads property.

Skift: Do you see social media platforms like Facebook being able to move further down toward the booking process?

Fiona Boyce: I think it’s two-fold. I think it’s a really effective marketing tool for awareness and branding and editorial, but we are seeing that with our spending with the dynamic Facebook ads and retargeted ads that we are actually receiving quite a few bookings from those as well. I think that we’re definitely making sure that our message is consistent but tailored. If we’re posting on our page, it’s going to be different copy than if we’re doing a dynamic, retargeted Facebook ad.

I would say that we have primarily used Facebook more for branding and awareness, but we are seeing guests booking through Facebook through those ads that we’re running. We’re starting to invest a bit more in Facebook and also in other social platforms like retargeted ads on Pinterest as well and promoted Tweets and also on Instagram and we’re seeing that our guests are booking through those channels. There’s definitely an increase in comparison with past years in people actually booking, so we are looking to test more and invest more.

Skift: What have you found most effective on Facebook and how are you tracking campaign effectiveness? Is it though A/B testing? Are you using other metrics to track what content works best for your brands?

Fiona Boyce: We do A/B testing. Actually, it’s more like A, B, C, and D. For example, with a boosted post, you can use several different photos. We experiment with Facebook dark posts as well. We definitely do a lot of testing, especially when we’re looking at something like a summer campaign. We’ll share a lot of different images, whether that be a food theme, or cocktail, fashion, cityscape, and we do test them and see what performs the best. We then go in halfway through our campaign and edit it to make sure we’re only sharing the type of visuals that are performing best.

We do a lot of testing, because our hotels are lifestyle hotels, there’s so much that we can talk about and so many different types of imagery and touch-ups that we can use. Our brand campaigns are typically pretty broad so it will be under a theme, but that theme can have a lot of different components to it so it gives us a lot of room to test and see what resonates. Also with imagery, there’s a lot of different communities where their followers engage with our content at a much higher level than others. We’ll do what amounts to something similar to a fashion shoot at one of our hotels and use that as the imagery for a summer campaign.

Skift: What are those dark posts that you mentioned?

Fiona Boyce: Those are posts that you can boost and they don’t show up in your timeline, so you’re not clogging your timeline with promoted posts, but you still can target that to a specific audience.

Skift: Are you experimenting with video at all? In terms of engagement with your audience, is it worth the investment?

Fiona Boyce: Yeah. We are experimenting with video and in 2017, it will be a big initiative for us. What’s really interesting about video is that we do see a much higher organic reach, but we don’t see as high of levels of engagement. Facebook is definitely trying to push video and who sees it, so we do get the lift there, but a lot of the time we’re seeing that the engagement is not as high as it may be on our best photo ads. Because the organic reach is higher and the reach is so low on Facebook normally, that is something that we want to definitely take advantage of. We’re looking at much shorter videos because that’s really what was performing well.

I just think that people are not totally watching videos on Facebook yet, but they’re seeing them more in their feed. We just need to figure out what is the actual content of the videos that will make people watch it and engage. That’s our big thing that we’re focusing on in 2017.

Skift: How different is creating advertising campaigns on Facebook versus Instagram?

Fiona Boyce: On Instagram, right now we’re mainly running pre-targeted carousel ads of images. We have done some promoted posts of more editorial content and it’s interesting because it performs really well with a lot of engagement on it, but we don’t really see any growth on our page. That promoted Instagram is being shown to a whole new audience that isn’t necessarily following us, but we don’t actually see a lot of people actually clicking over and then following our profile. I think there’s a disconnect right now between how promoted posts on Instagram interact in the feed, but we are doing those carousels of images with a book-now button and it performs okay. It’s not as powerful or lucrative as Facebook, right now.


The airline business is much different than lodging in terms of how Facebook is likely to be utilized.

We do not believe that broad brand building will be a key driver in an industry that is incredibly consolidated, especially in the U.S. We see the usefulness in dynamic retargeting, but will the airlines really pay much for this? The metas and OTAs give extremely low rates (if not free) to attract users to the sites to book lodging. Given the effectiveness of the existing relationships, there is not much incentive for either party to aggressively grow the relationship.

We do see messaging being increasingly important for the airlines. We will discuss this at greater length in the messaging section of the report, but at a high level, Facebook Messenger and WhatsApp can be effective tools to communicate with customers regarding delays, upgrades, flight status, luggage, etc.


Facebook has quietly added functionality in the past few years for local search. Conceptually, this could be very useful for in-destination activities. Facebook has the ability to search nearby places for restaurants, bars, hotels, shopping, etc. However, it’s a cumbersome process to find this.

In the app, first we have to click the “more” button at the bottom right of the screen.

Then, we can see the choices. In this example, we click restaurants. After that, we can click through the links and reviews and wind up on the restaurant’s website.

The current iteration of “places” is not that robust compared to something like TripAdvisor, OpenTable, or Yelp. We expect functionality to improve here, which could increase usage over time. Where we see greater potential is dynamic retargeting. A tour company for example could show ads to people staying at a certain destination based on demographics. For example, a family “pirate-themed” cruise in Hilton Head, South Carolina could dynamically target people with children staying on the island between the spring and summer months. We could see a series of three ads where the tour operator builds awareness when a person books a hotel in a location, then another in the weeks before to serve as a reminder, and finally one while the consumer is in-destination, potentially with a discounted call to action.


Messaging and bots are the latest buzzwords in the travel industry. While some believe that messaging platforms could be the way of the future for travel bookings, we expect instead that messaging will be used far more to enhance the customer experience. Facebook Messenger and its sister offering WhatsApp are by far the largest platforms in the world, each with over one billion users. Facebook has not reaped much in the way of direct financial gains from these platforms (yet), but we believe it’s just a matter of time before it does.


Facebook acquired WhatsApp for $19 billion back in February 2014. At the time, the price seemed high for a messaging app. However, WhatsApp is proving to be an incredibly important acquisition.

Facebook Messenger and WhatsApp dominate the global messaging market. The acquisition by Facebook was both offensive, bringing in a massive user base primarily outside of the U.S. into the Facebook ecosystem, and defensive, acquiring Messenger’s largest competitor before a company like Google did so. As the world becomes increasingly mobile-first, owning the largest messaging platforms globally is a unique competitive advantage for Facebook.


Simplistically, Messenger and WhatsApp are both messaging platforms that enable chat (one-on-one and group) along with voice and video calls for free. They use wifi when available and then use data otherwise. The only fees to consumers are indirect through data usage, but as data has become commoditized, global telecom players typically offer large data plans at very reasonable rates.

Both Messenger and WhatsApp facilitate international connections where there is no difference between calling or messaging your neighbor versus someone halfway across the planet. The platforms allow for sending photos and videos as well. Both platforms are contained outside of the Facebook app and now work on desktop as well as mobile.


Since Messenger was just an extension of Facebook, data sharing is expected. With WhatsApp, the entity is said to operate largely independently. In a 2014 interview with the Wall Street Journal[1], WhatsApp’s co-founder and CEO Jan Koum stated “Mark (Zuckerberg) proposed something very interesting two weeks ago, where he positioned it more as a partnership. He positioned it more as WhatsApp will stay completely independent and autonomous.”

Mr. Koum went on to say, “For us, it’s about staying independent. It’s about staying autonomous, as was discussed and announced during the deal. And it’s also about growing. We still have a mission. We still need to get to a billion users and then we need to get to two billion users. And we won’t stop until every single person on the planet has an affordable and reliable way to communicate with their friends and loved ones.”

In August of 2016, WhatsApp changed its privacy policy, which will allow more data sharing.[2] This is under review in parts of Europe, but it is clear that Facebook wants to use its messaging platforms as another source of data to make its ad targeting more effective.

We joined the Facebook family of companies in 2014. As part of the Facebook family of companies, WhatsApp receives information from, and shares information with, this family of companies. We may use the information we receive from them, and they may use the information we share with them, to help operate, provide, improve, understand, customize, support, and market our Services and their offerings.

This includes helping improve infrastructure and delivery systems, understanding how our Services or theirs are used, securing systems, and fighting spam, abuse, or infringement activities. Facebook and the other companies in the Facebook family also may use information from us to improve your experiences within their services such as making product suggestions (for example, of friends or connections, or of interesting content) and showing relevant offers and ads. However, your WhatsApp messages will not be shared onto Facebook for others to see. In fact, Facebook will not use your WhatsApp messages for any purpose other than to assist us in operating and providing our Services.



Facebook Messenger has tended to skew more toward developed markets and was initially housed within the Facebook site and then the app until being migrated to its own app in July 2014.[3] At Tech Crunch Disrupt in September 2016[4], Facebook’s head of Messenger, David Marcus, noted that part of the benefit of being its own app versus a mobile web experience is that it creates an ecosystem where everyone automatically has push notifications; Facebook does not automatically do this. This makes responses comparable to text messaging. They do still serve developing markets on mobile web where the user cannot use the app.

Facebook Messenger has dominated in North America and Australia while WhatsApp has been the leading chat platform in Latin America, Africa, parts of Europe and Asia (ex-China). The complementary audience is another key reason why Facebook wanted to buy WhatsApp.

As of May 2016, according to SimilarWeb[5], which used Android data in 187 countries, WhatsApp was the leading messaging app in 109 countries or 58 percent global country share. Facebook Messenger has a similar audience base, but is more concentrated in large, developed countries like the U.S. Messenger led in 49 countries or 26 percent country share. The combined Facebook-owned platforms are the leading app in 85 percent of the world. The only other app to have more than 10 countries was Viber, which is strong in Eastern Europe.


Facebook Messenger handles payments by working with entities like Visa, MasterCard, American Express, PayPal, Braintree, and Stripe. Messenger already has bots with an open API and shares data with Facebook.

WhatsApp does not utilize payments, enable bots, or open its API yet. It defaults to full encryption whereas Messenger has an option for secret conversations.


WeChat is the leading messaging platform in China. More than that, an astonishing 35% of time spent on mobile in China is on WeChat; 55% is on Tencent as a whole. Tencent also owns QQ, a messaging platform started from desktop. Both WeChat and QQ have close to 900 million active users.

Surprisingly, WhatsApp is still allowed in China, but that could change after WhatsApp’s new end-to-end encryption. That being said, WhatsApp is not widely used given its limited functionality compared to WeChat. Facebook Messenger is blocked with Facebook as a whole.


WeChat offers far more than messaging and voice and video calling. Functionality includes[6]:

  • WeChat Pay for online and in-store payments
  • Money transfers
  • Hailing and paying for taxis and Didi (the Uber of China)
  • News and media
  • Ordering food delivery
  • Paying bills
  • Managing a personal wealth fund
  • Donating to charity
  • Gaming

In January 2017, WeChat launched mini programs, which allow users to use other apps without downloading them. For context, if a U.S. Facebook Messenger user had this functionality, he could book an Uber without downloading the app. This has the potential to keep users on the WeChat platform even more than they currently are.

We do not expect WhatsApp or Messenger to become this type of platform given that users outside of China are already accustomed to using different leading apps for different functions, but illustrate how many other use cases there are outside of core chat.


The list below includes competitors to Facebook Messenger and WhatsApp like Line as well as discussing how the top tech companies touch messaging directly and indirectly.


Google is a bit of a puzzle in messaging. On one hand, it has some strong offerings. On the other, the multiple options make it confusing to figure out which platform to use.

The Allo app and the associated Google Assistant platform is actually quite useful for travel. It’s a kind of hybrid between Apple’s Siri and Google Search. A user can ask to find flights to London and the app brings back options. We can then click into the options where we either go to Google Flights or a pure Google Search. The messaging platform is not used to book per se, but more as an extension of the Google browser. For hotels, the process is comparable where the results can be a Google search call to the hotel or directions. Allo is text-based and similar to Facebook Messenger in that respect.

Duo is essentially a version of Apple’s FaceTime.

Google Voice offers both voice calling and text-based chat. It’s free in the U.S. and costs pennies per minute elsewhere. The benefit versus Allo and Duo is that a virtual phone number can be used so a user can use any device or computer to make a free call. The negative is that the platform is not very sleek and looks old compared to Google’s other options and peers like Facebook.

Hangout is inside Gmail and an app by itself for both iOS and Android. In all cases, it has text, voice, and video calling by phone number or Gmail address.

Messenger is Google’s Android text-only platform.[7]

Overall, Google has plenty of functionality and we would love to see the apps become more integrated into one Google app.


Apple’s iMessage elegance is in its simplicity. This is not a chat platform, but simply changes texts to iMessages.[8] While Apple is the dominant brand in mobile, Android is prevalent across so many brands that it has an 87% share of operating systems versus 13 percent for iOS as of November 2016, according to IDC.[9] This is one reason that we see Messenger and WhatsApp being more useful for chatbots in travel where Facebook is platform agnostic.

Siri is essentially an extremely powerful chatbot. Facebook’s messaging platforms are simply acting as the platform (for now) while Siri is the bot itself. In terms of breadth, Siri’s functionality is amazing. For domain-specific things (like travel), we find the answers given by company chatbots on Facebook platforms to be much better.


Amazon Alexa is the voice service that powers Echo. Here, the “bot” is voice-based. Several online travel companies have already started experimenting here. In November 2016, Expedia debuted an Alexa “skill” that enables customers to ask Alexa about their existing Expedia flight reservations, book a car rental, or inquire about their Expedia loyalty program numbers. Consumers can’t yet make a flight or hotel booking through the Expedia Alexa skill, but that’s in the works.[10] Earlier in the year, Priceline’s Kayak created an Alexa skill for checking flight, hotel and car-rental prices. Kayak’s Alexa feature also doesn’t enable travelers to book anything.

Alexa is constantly learning and improving, which should help build further use cases.

Amazon also offers Amazon Lex, which is a service for building bots across multiple platforms including the Facebook-owned ones.


Microsoft’s Skype, which it bought back in 2011 for over $8 billion[11], stated it had over 300 million monthly active users back in March 2016.[12] Skype has traditionally been known for its free Skype-to-Skype voice and video calls, but it also has text functionality. It started originally on desktop and now has a mobile app. Skype is a great tool for business and personal use, but we do not expect it to be prevalent in travel since free use requires both sides to use Skype rather than a mobile identifier.


Line is a Japanese messaging company that had its IPO in 2016[13] and is worth around $7 billion. The company gets 75 percent of its revenue from Japan. It offers the usual features that free messaging apps do (chat, voice, video) for its 217 million global monthly active users; 66 million in Japan and 101 million in Taiwan, Thailand, and Indonesia.[14] Line also has retail, food delivery, games, and stickers. In Japan, Line is installed on 80 percent of Android devices and each person averages 45 minutes of use per day.[15]


Japanese online retailer Rakuten acquired Viber in 2014 for $900 million.[16] As of December 2016, it had 260 million monthly active users.[17] It is prevalent more in Eastern Europe and has the standard functionality we expect to see across messaging, voice, and video.


As of February 2016, the app had 100 million monthly active users. The company is based in Berlin and founded by Pavel Durov, who has been called the Mark Zuckerberg of Russia. Durov went into self-imposed exile from Russia in 2014 after refusing to comply with the Russian government order to turn over data on Ukrainian VK users (VK is the Russian equivalent of Facebook that he created). Durov said he created the app in 2013 because he thought messaging apps were not very good at the time.[18]

Telegram emphasizes encryption, given the background of the founder. It’s only for messaging now, but is working on voice.

According to the company, “We believe in fast and secure messaging that is also 100 percent free. Commercial companies frequently face the need to compromise their values for financial gain. This is why we made Telegram a non-commercial project. Telegram is not intended to bring revenue; it will never sell ads or accept outside investment. It also cannot be sold. We’re not building a ‘user base,’ we are building a messenger for the people. Pavel Durov, who shares our vision, supplied Telegram with a generous donation through his Digital Fortress fund, so we have quite enough money for the time being. If Telegram runs out, we’ll invite our users to donate or add non-essential paid options.”


KaKao is based in South Korea launched in 2010. It has chat, voice, and video capabilities on its platform along with media, games, lifestyle (transportation), payments, and search. It had 49 million monthly active users as of September 2016 with 41 million in South Korea.[19]


Simplistically, chatbots facilitate chat conversations where the consumer engages with a computer rather than a person. According to Chatbots Magazine[20], “a chatbot is a service, powered by rules and sometimes artificial intelligence, that you interact with via a chat interface. The service could be any number of things, ranging from functional to fun, and it could live in any major chat product (Facebook Messenger, Slack, Telegram, text messages, etc.).”

There are two type of chatbots. The first is rules-based that can only respond to specific commands; it’s limited to what it’s programmed to do. The more powerful type is powered by machine learning. The bot can then understand basic language and gets continuously smarter as it has more conversations with people.


There’s a lot of excitement about messaging, voice, and artificial intelligence within the travel industry. Some believe that chat can become revolutionary where bookings, customer service, and in-destination tools all become chat-based across the major messaging platforms. Carried to its extreme, messaging would become the actual booking app. We are excited about the technology, but have a more tempered view of the likely use cases.


Customers booking travel on messaging as it’s first rolled out tend to be tech-savvy early adopters. For the broader travel consumer, chat is still largely a novelty. As more consumers try out the messaging not just for booking, but also for reservation changes, customer support, in-destination recommendations, etc., the chatbots will become more powerful in their ability to offer better responses. This in turn should increase usage, leading to a virtual feedback loop.

Messaging for communication is very appealing to the travel industry where consumers can use wifi when travelling abroad to place what amounts to phone calls and text messages. For any international traveler, this is especially appealing where roaming data charges are quite high. We expect online booking sites, hotels, and airlines to all utilize chat and bots to communicate directly with their customers.

Companies across the travel industry that best engage with their consumers will build increased brand loyalty, which in turn, helps build market share. We discuss the individual use cases by sub-segment in a later section.


We tested out the booking features for the large OTAs and metasearch sites. Our impression was that it is possible to get what we wanted, but simply using the app or the website was much easier to do. The online booking sites we spoke to acknowledged as much where chat remains in the early days of functionality.

On the hotel side, we tested asking Marriott, Hilton, Hyatt, and IHG by messaging on Facebook. Marriott, Hilton, and IHG did not respond and the top of the page said it could take an hour. For Hyatt, we were actually speaking to a human. We asked if he was a chatbot, but he informed us he is a person named John. The response time here was not great and it was far easier to search on Hyatt’s website than use the platform. We asked John if many people book through messaging and he let us know that they are not able to take credit card information through social media, so any bookings tend to be loyalty members with cards on file.

For airlines, we ran a similar test. United simply automatically replied with a link to their website. Delta did the same after twenty minutes. JetBlue took an hour and then said to call them. American did not reply. Internationally, Cathay did not reply, but Lufthansa quickly replied with a chatbot named Mildred. The interaction was smooth at first when we asked for flights to London from New York, but after it returned economy options, any other question asked received a “let’s continue in English” reply over and over again.


Messaging will most likely be monetized on two fronts. The first is through advertising. The company has started rolling this out, but this is very early days. Ad loads will remain quite low for the foreseeable future, but this is revenue that Facebook did not have before. We also expect that data will be used from messaging activity to better target ads on Facebook on Instagram.

The second path to monetization will come through a transaction fee on purchases done through Messenger or WhatsApp. The amount would likely be quite small and comparable to Apple Pay, which is facilitating the transaction, but is not a bank issuer nor a credit card company. Back in 2014, The Financial Times reported: “Bank chief executives fawned about the ‘exceptional customer experience’ and the ‘exciting move.’ They are also paying hard cash for the privilege of being involved: 15 cents of a $100 purchase will go to the iPhone maker, according to two people familiar with the terms of the agreement, which is not public. That is an unprecedented deal, giving Apple a share of the payments’ economics that rivals such as Google do not get for their services.”[21]

Below is a typical transaction where the total card usage fee paid by the merchant is the Merchant Discount Rate (MDR):

Example on a $100 purchase

  • Total MDR is $2.50, which first goes to the acquirer
  • Issuer gets $1.75 on the interchange fee from the acquirer
  • Network gets $0.25 from the acquirer
  • Acquirer keeps $0.50
  • Merchant gets $97.50
  • Customer pays $100

The parties involved are:

The issuer would be the financial institution on the credit card. This could be a bank like JP Morgan partnering with Visa (the network) or American Express, which operates the entire payment loop.

The acquirer receives a fee directly from the merchant. They take on the financial risk so the merchant does not have to along with authorizing, processing, clearing, and settling the transaction. They will route card information and the transaction details between the merchant and the network. The key functions provided to the merchants include banking, technology, and general customer support. PayPal is actually a specialized acquirer.

The networks include Visa and MasterCard. These are purely tech companies and do not take on financial risk. The networks take care of the actual infrastructure, clearing, settlement, authorization, fraud detection, analytics, etc.


Both on messaging and more broadly speaking, Facebook could facilitate online payments and handle the technology at the front end for a fee, perhaps in the 0.05% to 0.10% level.


In this case, Facebook could take the approach of charging a small commission for allowing transactions to be done on its platforms. Given the vast consumer base and engagement levels, we believe that either the merchant or the banks will offer a concession to increase purchase volumes.

Whatever the method is, we fully expect some transaction fee as we look into the next decade.


For the travel industry, specifically, we do not expect much in the way of transactions through messaging. Other vertical with simpler search parameters are more likely to have success on messaging for purchase. For example, asking a verbal “Where can I buy a window air conditioning unit?” would give a straightforward answer and a click-to-buy button, whereas travel is far less commoditized or uniform.


Our conversations with industry contacts indicate that advertisers do not yet use messaging data to dynamically retarget on Facebook and Instagram. We believe that will happen over time.

Additionally, we expect brands to engage with past and current customers by targeting them with deals in the messaging platform. For example, Kayak may message someone who booked a flight and offer hotel deals or in-destination options. The booking would likely move to Kayak, but the functionality makes it more likely that Kayak values the Facebook platform and increases ad spend there.


Even if booking through messaging does not catch on en masse, using messaging for customer service can increase brand loyalty. For online booking sites, a consumer may choose to only use that site for all travel when possible if the consumer is impressed with service after the booking. This is important, as the booking part of the transaction is largely driven by price and inventory. The largest companies dominate here. For them, maintaining share requires not only maintaining the best technology on search, but also having large inventory and the best customer support. Messaging and chatbots will be key parts of consumer engagement and support that further help distinguish the best brands. even enables chat between hotels and guests through Booking Messages on it its Pulse app.[22] The app also includes features such as a calendar, an overview of arrivals and departures, and special guest requests. According to David Vismans, Chief Product Officer at, “We know that 75 percent of our customers prefer self-service options to take care of simple requests. With the new Booking Messages interface, customers can literally customize and manage their trips in the palm of their hand. This technology takes away the worry of a busy front desk, different time zones, or friction of communication between people speaking different languages.”[23]

This is not part of Facebook Messenger, but is an example of the type of chat and product enhancements the top companies are using. This also builds loyalty with the hotel partners, which is important in building the best inventory and justifying rates that hotels often publicly complain about.

Below we provide commentary from Kayak, Skyscanner, and Momondo. We discussed all three companies at great length in our 2017 Outlook on Metasearch in Travel report. As previously noted, Priceline has now acquired both Kayak and Momondo where functionally it’s as if Kayak was the buyer.


Kayak operates 40 international sites in 20 languages, processing 1.5 billion queries offering air, hotel, cars, packages, and now, trains and vacation rentals (small offering as they test the market with 350,000 properties). The company has been the leader in the U.S. market and is starting to grow internationally. Priceline has taken a different approach with Kayak than Expedia has with Trivago as it’s not investing as heavily in advertising to fund revenue growth. The company likely still generates more revenue than Trivago and smaller peers, but is likely massively more profitable than most traditional metasearch competitors. Those peers are growing revenue nicely and will likely become profitable, but we believe Kayak is already the industry standard for monetization.

Giorgos Zacharia was appointed chief of technology in 2014, replacing Co-Founder Paul English. As chief product officer, Giorgos oversaw the company’s global technology strategy and product development for Kayak’s websites and mobile applications. He joined the company in 2008 as chief scientist. Prior to joining Kayak, he was co-founder and CTO of Open Ratings, Inc., a leading supply risk management provider. Giorgos also co-founded two machine learning-driven hedge funds, Stocknomics and Emporics Capital Management. He holds four math and computer science degrees from MIT, including a PhD, won five medals in the International Mathematical and Physics Olympiads, and was a Fulbright Scholar.

We discussed messaging and bots at great length with Giorgos. Below is a summary of that conversation.

Skift: How widely used has the chatbot been for bookings so far? Do you envision it being more for customer services and then growing in usage as bots as a whole improve? In what parts of the travel journey will messaging and bots be prevalent (planning, booking, in-destination)?

Giorgos sees users utilizing the platform for both booking and customer service, but his opinion is that it will be more about customer service. The users right now are mainly early adopters and that is why booking is fairly widely used. Kayak is investing in both use cases so Kayak will be there for all scenarios.

Messenger is also being used to communicate flight status, gate updates, luggage info, etc. Customers can ask Alexa those types of questions as well. Another benefit is the ability to use Messenger just on wifi when travelling internationally.

Skift: Can you walk us through the technology behind Kayak’s chatbot? How was it created, what factors are important in building responses, how many back-and-forth sentences between a bot and human are needed to drive a booking versus the steps in search on an app or website?

The technology behind the bot is machine learning. Kayak is training the bot on examples that Kayak learns from users where the bot learns how to respond to human questions. It’s important to have a large data set to learn from.

He believes that each back and forth correspondence is equivalent to 2.4 clicks. On the website, the fewer clicks needed to achieve a result drives better conversion where the rate of abandonment increases as the number of clicks rises. For bots, it’s even more important to be efficient as the platform is new and needs to build trust with consumers.

Giorgos noted that Facebook is also adding multiple choice click type actions, which reduces back and forth.

Skift: What is your view of third-party tech companies specializing in building bots? Do you guys partner with them or do everything in-house?

Kayak is a leading tech company, so does everything in house. He emphasized that when you have domain-specific expertise, it’s better to do it yourself, provided you have the technology horsepower, which of course Kayak and the Priceline Group possess where Kayak has the human expertise in the industry to drive the bot along with the technology and data to optimize it.

However, he agreed with our case that entities like boutique and independent hotels could be well-served by partnering with third parties.

Skift: What platforms other than Facebook Messenger is Kayak using for chat and bots?

He noted that WhatsApp has not opened its API yet, but he expects this to occur at some point. They have a Slack bot and an Amazon Alexa one and are working on iMessage interactions on Apple.

Skift: Broadly speaking, how do you see the evolution of messaging and bots impacting the travel industry? What time frame do you see for wider customer use and acceptance?

He said it’s hard to predict and joked that mobile happened quicker than anyone expected. If it happens quicker than expected, Kayak is a leader in the space and will be there.

Skift: What would cause chat to not be widely utilized? Is it on the tech side or consumer acceptance?

Giorgos stated that the technology capability is there, but delays would be more on customer interest.

Skift: When bookings are done through Messenger, do you pay Facebook a percent as you would if an ad led to a booking or is Messenger/WhatsApp still letting transactions proceed without taking fees?

He informed us that it us free, but does expect Facebook to monetize this. Once Facebook charges, if ROI is positive, Kayak will be there.

Skift: Does Kayak use data and conversations between the Kayak bot and a consumer to run dynamic ads on Facebook and Instagram? If not, do you see that playing out in the future?

This does not happen yet, but he can see that happening in the future.

Skift: What languages are available?

The current bot is in English, German, and Portuguese. The U.S. is the largest market based on total numbers, but as percent of users, the largest acceptance is in Brazil.

Skift: Will Kayak be part of all messaging platforms?

For now, it will focus on just the big ones and where the users are.


David Low created and now leads the Developer Advocacy program at Skyscanner. In his role, David looks after external developer relations, working to grow the evangelist developer community around Skyscanner and the Skyscanner API. David began with Skyscanner managing the company’s mobile products, including implementing a best practice standard. Prior to his Skyscanner life, David managed digital production at RBS and led product development at STV.

Skift: We’ve heard a lot about messaging and bots impacting the travel industry and but it seems like we are very much in the initial phases. How far away are we from where we see people open Messenger or WhatsApp and the experience is as seamless as it would be on the website or app? We played around with the Skyscanner bot and it’s great compared to other similar platforms, but when we compare it to the Skyscanner app or the Skyscanner website it seems it’s just much easier to not use the chat feature and stay on the app or website.

David Low: That’s a fair point. We’re still at the early stages of the bot and knowing what people want, what they’re going to ask and how we cater for it. We went through the original stage probably one year ago, so we were working on it a couple of months before Facebook announced the platform. We tried to recreate the basics of a booking funnel same as the app would have, just to get the basic five or six questions that really matter, but did not get into the deeper filtering part of it because we felt that would come later.

We’ve been listening to what people ask us for and often people didn’t want the booking funnel as such; they want to ask something a little bit more informational so they’re looking for, sort of advanced information before they get to a search. So, they’re starting to ask these things like, “What’s an average price for a room? What should I look to pay for something? When should I look to be booking this? Do I need to wait?” And things like that.

I’m not sure users know what they want to ask it yet, so we’re almost reacting to what we see coming in and tailoring it as we go so there’ll be a few changes coming up where you start to see a few experiments that we offer people different things and tell them what they can ask that they might not figure out they can do yet. So, we are still a while away from knowing really what it’s going to be, and even at the sort of conceptual level whether it’s going to end up being a transactional platform, soon that people willingly come and book a flight, through a messenger or if it’s a way to keep in touch with the user, to retain, and even to acquire them.

You’ve got people that might, just because they’ve seen the bot exists, come and ask some questions and that then introduces them to the rest of Skyscanner, so it’s more of an acquisition channel in the same way you might acquire them through Facebook advertising. There’s all sorts of things in play at the minute that are very experimental.

Skift: How early in the days do you think we are here? How many years away are we from messaging being widely used for different use cases?

David Low: It depends on how quickly the platforms react to the sort of things we find from people, but I wouldn’t say it’s five or 10 years away, but somewhere less than that. I spoke with my colleagues on the bot team at Skyscanner this morning about apps and how Skyscanner didn’t dive into apps early, but actually waited about two and a half years from the app store for iOS until it took its first bite. In this case, we’ve gone in really early, and not waited to see how the market matured before we dived in. So, we’re trying to figure out what our timescales look like and when we can expect, the behavior to change.

I suspect in 18 months it should have bottomed out by then. We’ll see what comes from the Far East as well because companies like WeChat will start to expand westwards and they’ve got a heavily transactional platform that is really built around payments. We’ll see if they’re able to influence what goes on in the west, because at the minute it’s a bit different where the bots in messenger are really informational and sort of a replacement for push notifications.

We haven’t got to the transformational bit, which is what happened in China, five years ago maybe. So it depends on how quickly the influence moves, but I wouldn’t expect it to be more than 18 to 24 months before that starts to figure itself out.

Skift:   Touching on China, I know Ctrip and Skyscanner have stated that Skyscanner will operate independently, but China has unique rules concerning restricting Facebook.  How would that work? Would Skyscanner still as an entity be able to use Facebook everywhere else outside of China and use messenger and then use WeChat in China? Or will it be different with a Chinese parent company?

David Low: I don’t really expect it to be any different. We are operating independently as we always did, so on that basis we’ll carry on working with everyone and the way we operate in China shouldn’t really change either because it’s always been a separate unit that works out there.

Skift:   Being part of Ctrip, we would expect China to grow in importance to Skyscanner. Do you see messaging being used on other messaging platforms like WeChat in China or Line in Japan?

David Low: So, to go back to your previous point, it’s still so early we’ve been a bit reluctant to do too many platforms at once, because it’s difficult to learn from those even in the United States. We’re concentrating on messenger because it’s got the volume and we have the relationship with Facebook from some other things we do and it’s been an easier way to experiment.

We just want to really understand the smaller audience and experiment that way because we think we’ll learn more from it.

Skift:   If Facebook starts creating their own kind of bot voice messenger platform like Apple Siri or Echo, how does that change how travel companies use these platforms? Would Skyscanner still want to build its own bots because it has the data on bookings and everything else? Or, given the massive budgets at Apple, Facebook, Google, and Amazon, would Skyscanner just piggyback on that technology and kind of design it through the big tech players?

David Low: That’s a good question so I sort of operate in my job at Skyscanner in both ways. So, in my own background it’s been b-to-b and trying to exploit our data and API into other places. So, that’s a combination of things we build ourselves and that’s why we work with others. We’ve looked at models where our data is used by the people in the way you describe, so we’re part of somebody else’s platform; we’re not against working that way.

In terms of Facebook, the way they’ve operated it’s not, a big thing yet, but they seem to have a goal, and messenger is almost informing things they do toward that goal of being a hybrid assistant with the Facebook M project, which hasn’t really got any massive volume or scope yet, but they seem to want to get this combination of human and bot interaction. This is something we’ve started to think about and it’s something the market will probably get to that people, if they’re not completely comfortable with bots, there are times that they have some human interaction to guide them through it, and over time there’ll be less of the human and more of the bot.

Google is seemingly trying to automate everything and Alexa is trying to create a voice assistant. If Facebook gets that mixture of human and bot interaction right, that’s really where it starts to take off, because the trust will be there.

We would love to be involved in that kind of thing. It’s early days, and when that gets opened out to companies like us to be part of but we’d certainly like to be in there whether it’s us building something within a platform like we’ve done for Alexa in the last year, year and a half, or if we offer data into another service to improve it because of what we know and now we can feed into somebody else’s platform. I don’t think we’d be choosy about it. I think to your previous question, we think Facebook with that hybrid model is the differentiator in the market.

Skift: Can you give us data on traffic volume? What’s the uptick in terms of customer adoption? We know it’s still early stages but it would be great to get a sense of what you guys are thinking in terms of how wide scale this could get.

David Low: I can’t give any specific numbers, but, the volume’s better than we expected, to say that much. We launched it around the 15th of May last year. Volume and the conversion from the traffic is good. It’s better than we expected and it compares, very well against our mobile apps to use a comparative. People that come through the bot actually make a booking and that’s something we didn’t expect quickly and that’s maintained.

We’re not too worried about the absolute numbers of users even though they’re better than we thought. What matters is what they actually do once they’re in there. The sort of things they ask, how long it takes them to get from zero, sort of onboarded to the bot through to actually getting what they were looking for, so an exit from the bot, if you like, to complete.

We started with quite lengthy conversations where there was a back-and-forth friendly what’s this like, where are you going, when are you coming back, to people actually asking quite long and complicated things. Some of those more advanced long-term things, we maybe can’t answer yet, but we’re working to.

That’s the sort of benchmarks we look for cause even though these bots are conversational, for users to get used to them, the conversations really need to get shorter over time. We’ve certainly seen that as people use the bots over time, the conversations have gone from maybe eight, 10, or 12 interactions, down to two or three and they’re able to get what they want. The ideal for me would be to get what they want in one sentence so they could ask us something very complicated and we give them the answer that suited them and that’s where we’re aiming to get to.

Skift: How far along are we for Siri or Echo to help users actually locate flights or hotels? How might that work on the branding side in terms of white-labeling that kind of experience?

David Low: Siri for the first time, opened itself up to third-party integrations last year. Things like taxis were first and then eventually it was opened up to restaurant bookings. It’s typical of Apple that they’ve gone quite niche, but tried to do things really well.

For Facebook, Alexa in particular is probably a better comparison where you can build absolutely anything and what they’ve found is that people developed things and the quality started off pretty low because you have to train Alexa to understand what people say. Siri’s gone for a very narrow domain of things that it will let apps answer for, and really optimized them so if you ask for a taxi you’ll get a taxi and not something completely different. We’ll see how that works, but it was interesting that travel was left out of that.

Maybe it’s too difficult because, the thing about voice search in particular is dealing with the number of places people might ask for and the training involved in that is immense.

We’ve trained something like 300,000 phrases through Alexa to get it to understand geography and there’s way more to go just in English. So to get that to have aligned with it, it could be a long, long time before there’s sort of a global utility for these things, cause Apple will want to deal in every market straight away if they can, and actually training it to do that is beyond even that size of company. It’s really further away than you think for those technical reasons. Just making it useful enough for Apple’s perfectionists, is something that’s going to take them years to do so we’ll see but it certainly seems some ways off yet.

Skift: Everybody’s coming out with their own bots these days. How does that incorporate with the Siris and the Alexas of the world?

David Low: What we’ve chosen to do is try and run one single platform that answers natural languages questions. And then you’ve got a layer above that, when you deal with Facebook everything’s tight so you get the raw text that the user said and you interpret it yourself and answer it. With Alexa, they put the voice understanding in front of that so, you then have the job with Alexa specifically of training it to hear and understand certain things. You supply Alexa with a phrase such as find a flight from London to New York tomorrow, and you teach it which things are the keywords to extract from that and what they mean so the order to and from and from and to and the different things people can say.

To use Alexa as an example you have to do that specific training yourself; there is some help from Amazon to do that but you’re left to do most of it yourself and that creates a bit of competition between people like us and other travel companies to be better listeners. And then you’ve got stores that rate the apps, the public rates them the same as you rate Apple iOS apps and Android apps.

What we’re actually doing is trying to build something that’s a bit more flexible beyond bots potentially to other uses where we’re supplied information in any way, and able to give the traveler what they wanted. We thought we’d be better putting these blocks in now rather than just build a bot, which a lot of people have done. That’s fairly limited in what it’d be able to do if the market changed; we’ve tried to build something a bit more flexible and long term.


Below is a continuation from our broader conversation on digital advertising from earlier in the report.

Skift: Have you started using messenger for customer service or to facilitate bookings on any of the platforms?

Tore Pein Jensen: Yes. We launched a messenger bot and I’ve played around with that. It’s part of playing with the future that is coming. It’s part of the core product. We strongly believe that this is the future, there’s going to be so much happening on bots and artificial intelligence. It will change the way that we interact with the outside, the agent, and how people search on our site.

Skift: If we start looking out five years from now, and you see messaging and you see bots, do you think that’s going to be the case where people see your ads on Facebook and then they’re using messenger and WhatsApp to then actually facilitate a booking on Momondo in meaningful ways? How far away do you think we are to bots and messaging actually being a potentially game changing part of the booking process?

Tore Pein Jensen: I don’t know. In these terms, pretty far away. We need to divide the booking, inspiration, and general search. Bringing awareness to different destinations, I think that will seem pretty fast coming, and people will demand that service from us. Of course, we need to be there.

It’s a bit longer out in terms of years before we will see actual bookings in massive scale. Of course, there will always be nerds like me, that will try everything that is new out there. I think that will take two years before we are there. I envision long-term speech search with all that’s going on there, that will change the demands for how we should have searches.

Hugo Burge: The promise of travel search is to make travel search more seamless and simpler. I think you’d be hard to argue that we’re anywhere near chatbots revolutionizing the travel industry. It’s a very exciting area. The promise of being able to make travel bookings in a seamless way is a beautiful idea, but I think we’re a long, long way from that. It’s very early days in testing and learning.

We’ve got our chat bot on Facebook. It’s great learning for metasearch. It’s a different question. We don’t facilitate bookings. This is not an area that we see overnight being revolutionized. It’s an area we watch carefully and I think it’s an area where, as you’d expect, some startups are saying it’s a revolution. I don’t believe it’s a revolution, I think it’s an evolution. I think they’re specific use cases, where it could and can be valuable.

I think it’s interesting that companies like Lola, which is starting with human beings, and with very specific use cases, and just within frequent business travel. Because travel is incredibly complicated when you look across the spectrum of what people are searching for, and how you need to personalize responses.

This is a discussion point, we’re very early in our planning for this, but we believe that personalization is a key foundation and bedrock of good chat. Because ultimately in order to respond to queries to facilitate good responses, you need to understand who you’re talking to and give intelligent responses. I think we’re at the very early stages of that. Personalization is an important building block, and something that we see is something that we would work on first to get right.

Skift: Do you have any data on bot usage you can share?

Hugo Burge: For the Cheapflights Chat FB, which launched back in June 2016, there have been 20,000 active chat threads since launch converting into flight searches at 150 percent. Traffic generated from the campaign we ran when we launched Cheapflights Chat carried a conversion rate that is 38 times higher than traffic from Cheapflights Facebook travel content posts. This level of adoption for such a new technology feature, and the results demonstrated so far, place Cheapflights at the forefront of Facebook Messenger bot usage. But most importantly, people are using it to perform travel searches (on average two to three searches per user) and it’s proving a really useful tool for users. Although exciting, we don’t see Bots becoming the dominant search tool for some time. Right now, they’re a great addition to the canon and are showing potential.


It’s unlikely that messaging ever becomes a meaningful booking tool for hotels. Driving direct bookings is difficult enough via websites and apps, so we see messaging here more as a tool that helps improve the guest experience. The ability to reach out to guests directly makes Messenger and WhatsApp attractive tools. The hotel industry should use these platforms to answer customer inquiries both before and during the trip. Additionally, messaging can be used to up-sell or push ancillary services. For example, the hotel may have unfilled suites and offer a guest a Facebook Message an hour before checking in that she can upgrade to a suite for a fee. During the trip, it could recommend restaurants at the resort or activities. Hotels should have chatbots and humans to engage with guests.


Paolo specializes in e-commerce and electronic distribution with a focus on digital marketing, business intelligence, revenue management and CRM at Two Roads Hospitality, which is an international lifestyle hotel company created through the newly-merged Commune and Destination Hotels. Comprised of Joie de Vivre Hotels, Thompson Hotels, Destination Hotels, tommie, and Alila Hotels & Resorts, the company is the leading operator of independent and lifestyle hotels with more than 95 properties in eight countries.

Skift: What percentage of advertising goes into digital versus television versus print?

Paolo Torchio: We are almost 100 percent digital. There’s a couple of our properties doing some billboards, that type of thing, so I’m going to say 100 percent is probably a stretch, there’s probably some other stuff going on, but it’s very little.

I think you’ll see that percent is consistent with normal independent type hotels out there. Obviously for branded properties, they’re going to have the television budget, but we don’t. We don’t do much, if anything, at that level.

Skift: Within digital, traditionally how much goes to Google versus Facebook versus other channels, and do you see that changing as Facebook has improved their travel offerings?

Paolo Torchio: It is probably 80 percent to Google and then MSN, then Yahoo, some of the other players like that who are targeting with other components. Of the 80-20 split, Facebook is in there and is growing. As they evolve and do some pretty interesting things, more of the budget is going to be shifted that way.

As you look at smaller properties, there’s an expectation on return and that it has to be transactional, and Facebook is moving that way. We’re starting to see more reservation type activity coming from that advertising. As that gets stronger, I think you’ll see the percentage will spread to Facebook much more quickly. It’s almost at the tipping point where we will see if Facebook can become a transactional vehicle? It’s getting there, but it’s not quite there yet.

Skift: Do you think that switch comes from Google or comes from everything else?

Paolo Torchio: I’d say we’ve moved past experimental phase with Facebook so it’s starting to come from the Google side as well. This year’s budgets have got a line item for Facebook whereas last year did not.

Skift:   How do you see Facebook progressing in participating in the booking part of the process?

Paolo Torchio: We are starting to see activity and are actually starting to track bookings back through the Facebook referrals.

The transactional piece could really evolve quickly if the adoption of the messenger side of things and e-commerce within the messenger platform takes hold. That’s not just going to be a travel thing, as soon as that takes hold, the messenger platform becomes seen as a place to conduct business. That’s when you’re going to see the transactional aspect jump significantly.

When you look at the whole guest journey, when you look at that moving down the funnel touching different pieces of media or technology, a lot of that could start happening in Facebook itself. You go from a high funnel awareness within the Facebook platform, you move that to the messenger platform and that could become very transactional. Those messages could be for customer service or communication, but also could be transactional. You’re touching multiple points of the journey with the Facebook platform.

Skift: At some point, it seems like Facebook is going to monetize the messaging platform. As people start getting more comfortable with messaging and bots get better at answering questions, do people actually start booking on messaging platforms?

Paolo Torchio: What we’re describing now is that Facebook has become a platform. It is no longer just a social network, but a platform that is offering all these different technology services.

Facebook is evolving in much the same way that Amazon started selling books, but then it became a storefront and you weren’t buying from Amazon anymore. It then became a major mall and you would buy the technology or e-commerce within Amazon. Facebook is becoming a platform of sorts, not unlike Google and maybe others.

If they partner with payment processors and are hooked up with banks, monetization could come from the credit card side of it or through the relationships with the bank. It remains to be seen which direction they go, but I think this all goes back to the question of when does it become transactional.

Transactions are probably more likely in the chat side than within Facebook, which I think will stay a referrer. But it could refer into chats and then that’s where the transactional is probably going to take off faster.

Skift: That’s really interesting. Do you ever see messaging potentially catching on in the U.S. and Europe to be as popular for transactions as it is in China?

Paolo Torchio: I think it will get close, but it will probably never have the same acceptance. There’s a cross-section of the market that is only using their mobile device, and WeChat became this platform to connect e-commerce in China. So, I think in the North American markets we have too many devices.

If you look at the average consumer in China doing 80 percent of e-commerce on a mobile device and 80 percent of that is on WeChat, we’ll never get there in North America. But it will increase with the adoption of the messenger platform for commerce. It’s already starting to be more prevalent that you do customer service through messenger platforms. Commerce comes right after that, and as the consumer gets used to doing it there, hotels will follow.

Skift: Does Two Roads have chat and bots set up with Messenger, WhatsApp, or other platforms yet?

Paolo Torchio: We are doing some things with stand-alone apps, your Checkmate, your Alice, we’re testing a lot of those platforms, but haven’t moved into doing anything yet. There is a number of players in travel looking at solutions on Facebook messenger. We haven’t started doing anything there yet but it’s coming. We’re in the design phase of a type of affinity program.

Messenger platforms are going to play a big part of that and we want to move the conversation into that platform.

I think another element where hospitality is uniquely positioned is the concierge-type engagement. That’s also where a Facebook plays into that whole spectrum. You can think of a scenario with a high-funnel ad generating interest in a property, the property gets booked whether it’s on a Facebook platform or not, but then you move the consumer back to the Facebook platform, but now on the messaging side for guest service.

There’s an element of retail to that as well, there’s a great deal of talking to the concierge and that’s moved onto the messaging platform. That’s kind of the vision we have for leveraging those platforms, and it will be tied to our affinity program to build brand awareness across the portfolio.

Skift: As you build that out, do you see this being done in-house or using third-party specialist companies.

Paolo Torchio: We would definitely work with someone; we wouldn’t build this in-house.

With the size and scale of our company, we don’t develop technology, we buy technology and we output programs on top. I mentioned it’s going to be part of our loyalty, so it’s going to be important to find a white-label solution.

Skift: Going back to the core Facebook platform, they rolled out Dynamic Ads for Travel for hotels, what are your thoughts on this, how effective it is, what needs to be improved, what you like, what you don’t like?

Paolo Torchio: The real value there is the ability to focus on the audience. You’ve got intent from many different sides, what you’re looking at, what you’re clicking on, the network, your location.

I think it’s going to be a very effective way to spend money looking at the capabilities to do some dynamic pricing and group availability, leveraging API’s back to our information systems. That’s going to be critical. If we could do that in real time, that’s going to make it very effective. I think that’s almost there on the platform.

Skift: Is there anything else that you think is important to note?

Paolo Torchio: I thought when this interview request first came through, the Facebook platform is interesting as there is an evolution of a site that is moving away from being only a social network fueled by advertising.

Then there is this whole messaging piece, where it’s gaining momentum. For me, it’s playing in many phases of the journey when you walk through the funnel, but the guest service thing is the really interesting aspect that could become that platform where there is a lot of engagement happening. I think it’s a pretty interesting topic to be covering right now.


Similar to hotels, we do not expect booking to be a meaningful driver here. Instead, we see great value for messaging to communicate flight status, gates for baggage, upgrade options, in-flight requests, and as a tool for the airlines to engage customers ahead of and after the trip. Given the unpleasantness of air travel in many cases, airlines that go the extra step can create more brand loyalty.


Third-party bot technology providers will provide a valuable resource for the travel industry. This is especially true for the hotel industry where there are countless small boutiques that have a great brand and customer service, but lack the technology in-house that a Priceline Group would have to create its own bots.


Headquartered in Silicon Valley, Gupshup is the world’s most advanced bot and messaging platform. It enables developers to quickly and easily build, test, deploy, and manage chatbots across all messaging channels. Gupshup’s messaging platform is used by over 30,000 businesses, leading companies such as FlipKart, OLA, Facebook, Twitter, ICICI, HDFC and ZeeTV. Gupshup’s platform handles over four billion messages per month and over 150 billion cumulative. Gupshup also developed a smart-messaging app, Teamchat, which introduced patent-pending “smart” messages in 2014, only now being offered by other messaging apps. Gupshup’s bot platform provides tools for the entire bot lifecycle enabling developers to quickly and easily build, test, deploy, monitor, and track bots.

Gupshup is founded by Mr. Beerud Sheth, a serial entrepreneur, who previously founded Elance. Gupshup is funded by investors such as Charles River Ventures, Helion Venture Partners, and Globespan Capital Partners.

Skift: Can you walk us through your background, how you founded Gupshup and what the company does with messaging and bots?

Beerud Sheth: There’s a lot of buzz around bots, but the way I see it is that bots represent a major technology paradigm shift. What I mean by that is if you look at the last 30 to 40 years in the tech industry, every decade or so, the very way in which we use computers changed.

In the mid-80s, when the PC was born, the primary method for it was desktop. We had desktop OS and desktop applications. In the mid-90s, we switched to a browser and website model. In the mid-2000s, we then switched to a smartphone and mobile apps kind of model. Now, a decade later, we’re on the verge of yet another major shift, this time to messaging and bots.

There are many reasons why this is happening. People hate to download too many apps. The number one app by far is messaging. You have three to four billion people using messaging apps every day, dozens of times per day.

Bots are just the latest reincarnation of the website or a mobile app. They operate within the messaging app. You can think of the messaging app as a new browser and the bot as a new website.

Gupshup is one of the largest, one of the world’s largest cloud messaging platforms for SMS. We handle 64 billion text messages per month, because we realize that messaging was the only way to reach billions of devices out there.

Over the last few months, existing messaging apps also opened up their APIs so we just expanded our platform to support not just SMS or team chat, but also almost 20 other messaging channels. We also have connected with the voice messaging bots and just announced a partnership with Google Home a few of weeks ago.

Skift: On the travel side, we’ve heard Expedia and Priceline talking about messaging and bots and AI and how it’s going to be increasingly important. Do the travel companies use all the different messaging platforms or will they only plug into a few like Facebook messenger or WhatsApp?

Beerud Sheth: In general, the way any brand can leverage this bot phenomena, is they need to be on every platform. The bot encapsulates all the capabilities of that brand. For a travel brand, all the travel related activities would be enabled through the bot. It will be booking, making changes, confirmations, checking in, etc. I like to think of a bot as the brain, which can manifest itself in a lot of different places like Facebook Messenger, Slack, or Skype. It may also be available in international messaging apps and Snapchat, because users, especially for travel, may come from anywhere in the world.

Then you have voice activated bots. They may be on Amazon or Google where the user could just talk, right, say, “Hey, check me in for my flight? Or can you cancel? What’s the next flight available?” etc. Going further, you can even input the bots inside your website. You can have a little chat window that users can just chat with. It could even be inside your own app. You can have an app bot as well.

The point of this is that these interfaces are conversational. They’re very, very simple. Compare these with structure interfaces that you had on websites and on apps where you take screen number three, button number four, and tab number two. It’s a very structured interface. Sometimes as a user you’re just like, “When’s the next flight?” You expect the bot to know who you are, where you’re going, what booking you already have, and you want to know the next flight as opposed to go through this whole search process all over again. I think that’s the way to think about it. You’ve built a brain. You built a logic if you will. Then you deploy it across many different messaging, text and voice messaging channels.

Skift: Our view has been that initially you’ll probably get traction with the bots on messaging for customer service and potentially voice, but it may take until 2020 or beyond where people really start using it for booking. Do you agree with that?

Beerud Sheth: Actually, I think this phenomenon is going to explode much faster than people think, not just in travel but across the board. We’ve seen these kinds of shifts before when we went from desktop to web and then from web to mobile. It takes a little bit of time. The earliest versions are simple and primitive, but then they get better and people start using it. Then it becomes an inflection point. Even more importantly is the fact that bots are actually literally invisible technology. There is zero behavior change.

There is zero even awareness on the consumer side that they’re now using a bot. What I mean by that you neither know nor care whether it was a human or a bot on the other side that’s responded to a query. It becomes easy and simple and feels human and conversational. I think that’s why I’m very optimistic that this is going to happen in a huge way.

You don’t have to have a 1,000-person call center. You’re just running it through software. I think this is going to be huge and going to happen very rapidly.

Skift: What does the transition look like between the chatbot and a human agent for more complicated requests or when a customer experience session may not go the way the chatbot expects?

Beerud Sheth: What we do with many of the companies we work with is that people think of it as a dial that you can gradually turn on incrementally. You insert a bot between the user and the human agent. The bot can in the beginning just watch the messages flowing by. It may handle the simplest queries. At some companies, 25 percent of the queries are simple things like I lost my password or how do I do something. Those questions are the low-hanging fruit and can be automated fairly quickly. With more complicated questions, they’re going to the human agents. Then gradually, as the bots get smarter, they can start taking on more queries. By the way, it may never, handle every complex use case. That’s fine. Even if you get 20, 30, 50 percent of the queries automated that’s a big, big deal in terms of responsiveness, in terms of customer satisfaction, in terms of cost on the developer’s side. It’s enough where we may have this hybrid model going forward where maybe it’s 50, 75, 80 percent automated and maybe humans are freer to actually handle those complex queries that only remain.

The only other point I’ll make is as consumers, we evolve in interesting ways. I think about how Google has trained us how to do a search query. With a bot, I may just say hey modify booking, what options available, and go through that bot. I think this will evolve. It is not going to be very challenging in terms of how the adoption happens there.

Skift: When I look at Facebook, they paid a lot of money for WhatsApp and investing in both of its messaging platforms. How do you think they’ll eventually monetize it?

Beerud Sheth: Facebook probably is the most mature in terms of learning how to monetize social media. If you think about the social network side of it, Facebook did a great job of enabling people to talk to businesses and getting paid for it, but without feeling very intrusive. I think that’s the same thing here with messengers and bots. They’re enabling consumers to talk to businesses. So far, they haven’t said anything about how they’re going to monetize. I’m fairly certain it’s not going to be a message fee. I also think it’s unlikely it’s going to be a transaction fee though they might enable payments.

I think they know enough not to kill the golden goose before it even flourishes or gets to a mature point. I think they’ll do reasonable things. For digital content, it may be some sort of revenue share or payment processing. There could be some processing fee, things like that.

The key, the bigger prize for them, is to make sure all the engagements stay on their platform. I think it’s the same thing with bots. They’ll probably wait at least a few years before they start looking to it apart from some obvious monetization opportunities like payment processing and so on.

Skift: Turning back to the users, the big OTAs and the hotels, do the bigger companies create their own bots with in-house tech experts or do they use firms like yourself? How does that typically work?

Beerud Sheth: Like any new technology, there’s a learning curve involved. Fundamentally, Gupshup is a bot platform. We have a tool for developers that makes it very easy and quick to build advanced bots. As we scale up, we want to be a platform company. We don’t want to be a services company forever. However, in the early stages, because we have a ton of experience and expertise building bots, we do help companies that don’t want to leverage their resources. In fact, with some companies they use us to build their basic version of the bot and they take it in-house. That way, we can break them off the learning curve very quickly. We can get them started with best practices.

I think you’re seeing a combination of all of this. I think you’ll see both. Just like you have web development agencies, there will be bot agencies popping up over time. If a company had that skill set in house, I’m sure they’ll try to keep it inside because it will become a core competency. There’s a lot of components here around actual language processing, artificial intelligence, and the whole consumer experience that will be very core for certain companies. Over time, it will be in-house. I think in the short-run, they will use companies and we are happy to help as well.

Skift: When you have many messaging platforms, is it very different to program and code across the different systems or is it pretty fluid to go from one to another?

Beerud Sheth: It depends on how you do it. If you literally go down to the actual API, it can actually become very painful because each of them may have an arcane, specific requirement. They each can do it slightly differently. That’s where a platform like Gupshup comes in.

We are creating this layer of cross-platform API. You only build a bot once on Gupshup and it will automatically work on 20 different channels without you having to do anything at all. It’s not just text by the way; it’s even voice.

Skift: As a cross-channel platform, are you looking at the different messaging technologies? Are you following the traffic in terms of where the audience is, in terms of WhatsApp, Facebook Messenger, and do you foresee any particular messaging platforms that are a little bit more compatible with the chatbot paradigm?

Beerud Sheth: Right, so great question, I think even though among messaging app companies there are different levels of awareness and capabilities.

WhatsApp has not even opened APIs yet. You can’t really have any chatbots in WhatsApp just yet though they’ve announced that they’re going to do it soon. Facebook Messenger has been very aggressive. They’ve actually allocated a whole bunch of smart engineers, so you know you can render rich messages inside the messaging app.

Facebook is certainly very good. They’ve been the early leader. Slack is great for bots as well. In fact, they launched even before Facebook. It’s an Enterprise app. Telegram was actually also very early and I think Skype is catching up now.

For voice, you have Google Home and Amazon Alexa. Google’s going to put it on every Android device. That’s going to give billions of users voice capability as well. If you look at the growing numbers, the early growth has been around Facebook Messenger primarily.

Skift: Is there anything else do you think that’s important to note that we have not covered?

Beerud Sheth: No, but the rabbit hole goes pretty deep. I can chat for hours more about it. I think travel is going to be transformed in a very significant way. Travel has always been an early adopter segment. It was for web and for mobile, so I expect that to be the case with bots as well. The use cases are perfect. I think it can be truly transformative and can dramatically improve consumer experiences.


Preethi is a software and UI (user interface) engineer focusing on defining, designing, and building user interfaces for bots for, which specializes in bots for messaging across platforms and industries. She’s passionate about conversation interfaces and the future of human versus bot interactions, and enjoys the challenge of turning complex UI requirements into clean, intuitive, and beautiful user interfaces.

She was previously a front-end & API engineer at Coinbase, a digital currency platform, where she helped architect and rebuild their front-end platform in React.js. Prior to that, she was a partner at the top venture capital firm Andreessen Horowitz and an Investment Banker at Goldman Sachs.

Skift: Can you discuss your background, in general, and how you guys are different than some of the other bot platforms out there?

Preethi Kasireddy: I’m mostly a front-end UI (user interface) engineer, and I’ve been a generalist, but recently became really interested in the bot space, because of the UI challenges there. The usability challenges are very interesting, and it’s a very undefined space. It feels like Facebook is the only giant monolith currently, who’s taking over, or is at least currently just building Facebook bots, but I think the potential is much bigger than that.

There will be bots on websites, there will be bots in mobile apps, there will be voice bots, everything. The challenges there are very interesting, which is what excites me, and why I’m working with is focused on building bots for clients, and what differentiates us is that we do everything.

A lot of people, when they’re building bots, they don’t really know what they want, or what it should do, or what the experience should be like. They don’t really know how to build dialogs, so do the whole thing and provide the full experience back to them. That’s why we’ve been able to have a good number of notable clients work with us, like the Sacramento Kings, the Denver Broncos, and we have the Arizona Cardinals lined up. We have a grocery store called Publix, that’s doing a POC (proof of concept) as well as Home Depot and a bunch of other clients in the pipeline.

The reason we’ve been able to get these big brands is because they’ve never really built a bot before. By providing the full service, we can differentiate there, and give them what they’re looking for.

Skift: We were playing around with some of the bots in travel. It seems really early in the ability to use it for booking. It can be used to book and search, but it’s not the most pleasant experience versus just going to the website or app. How far away do you think we are for bots and chat being used for booking in travel, versus more of a customer service tool?

Preethi Kasireddy: You’re totally right. Everything is really early. When I first jumped into this, for the amount of hype and talk there is, when I first saw the bots themselves, and actually touched them and played with them, I was pretty shocked at how not sophisticated they were, but I think it’s just a matter of time.

I think people have started with the low hanging fruit and the easy use cases. The more interesting use cases, and the more advanced use cases, I think are probably at least another year or two before you start to see very well developed use cases, and it’s kind of like a chicken and egg problem, because the experience right now isn’t that amazing, so customers aren’t necessarily thrilled to use bots, so people aren’t necessarily sure if customers want to use the bots.

I think as bigger brands take a lead on this, and start to build more sophisticated bots, then you’ll see customers using them more, then us having more data for the customers, to be able to say, “Okay, this bot did this incorrectly, and it should have done this,” and using that data to feed back, and improve the quality of the bot.

Right now, they’re not sophisticated enough, people aren’t using it, so there’s not enough data to improve the bots. That whole cycle of building up that data set and getting users to adopt, I think will take at least another one, two, maybe even three years.

Skift:   For the non-engineers, such as ourselves, can you walk us through how you create the bots, how they’re deployed, and how they’re used across different messaging platforms?

Preethi Kasireddy: Sure. First, we create a scope for it. What is the specific, or a few specific use cases that we want the bot to do? On a high level, we define what features the bot needs to have, and based on those we decide what the dataset is we need. Then, based on the dataset, we organize the data and build the dialog. This is one of the most important and critical parts of it, even before implementing it. Building on all of the conversational trees, the flow, the different directions it can go in, how to reroute to the main dialog, or jump to different dialogs from one dialog to the next, this is the part that we’re currently working on, or trying to improve on the most, because this is where it starts to feel more human.

We build out the dialogs in these easy diagrams and Excel sheets, then once that’s all done, then engineers go and implement these. We’re currently using Microsoft’s Bots Framework, which provides a lot of very, very useful tools for building things like dialogs, and looking across different platforms so you can build one back end, and you can work across two, three, four different platforms, like Messenger, and Slack, or just a regular web interface and so forth.

Once that’s done, you start testing it, and improving it over time, and also plugging in an analytics solution, so that you can determine where the bot is responding incorrectly, and use that data to feed it back in and improve it over time, and maintain the bot.

Skift: For bots to improve, is it about machine learning and AI, or more so about humans, through trial and error, managing it, or some combination?

Preethi Kasireddy: Yeah. Right now, we’re using the LUIS (Language Understanding Intelligent Service[24]) API. LUIS is a Microsoft framework, that’s the AI component of it. What we do is we train LUIS to understand the customer intent. Based on what someone says, we train it to understand these intents, and based on those intents, we then match it to a certain dialog.

That’s the AI component, or the natural language processing AI component that we are currently improving on. In terms of using machine learning to actually basically have the bots learn and improve on their own, that is something we haven’t gotten to yet.

Over time, as you have more data, the pipe dream is that the bot can use its own data to improve and learn on its own, and that kind of AI is not something that we’ve gotten to, or I don’t think we’re even near there, yet, but right now the AI is more around actually being able to understand what the user is saying, the natural language processing part of the AI.

Skift: To get to the point where people in the travel industry would use a bot, and a messaging platform for bookings, do you think the technology capability is there, and it’s just the usage or do we need more on AI for that to be successful?

Preethi Kasireddy: That’s a good question, because it feels like the technology is getting better by the day, but the growth, or the sophistication of the product hasn’t really stepped up or grown to the same amount.

I think the technology is definitely there to do more than we’re doing right now. That’s my opinion. I think it’s a matter of finding really good use cases and actually implementing it properly. I just think the industry’s so new that there’s not that many engineers who know how to do this, there’s not many UI engineers who know how to do this, so people are just creating these hacky bots. They’re not really moving off anything beyond Facebook and Slack. Those seem to be the prominent platforms.

I think that just shows you that people are starting. They’re working on the low hanging fruit, but they haven’t figured out the bigger picture of these bots yet. Voice is a whole other story. That is probably the true future of bots, beyond the interface.

Skift: You mentioned voice is kind of the future of bots, and Siri and all the other voice activated search features have been around for a while, and they’re getting more and more sophisticated. What is ultimately the fundamental difference there, between a Siri and a bot that might leverage voice on the technology side?

Preethi Kasireddy: Siri is an example of bot. There’s no difference. It’s just that I think Siri works fine for very general use cases, but when you get to a specific brand, or a specific company, or customer service, or any of that, it can’t do any of that.

I think the difference is that there has to be much more data to be able for it to be useful, and put it in context, and basically every context you could think of.

I think Amazon Echo is probably the furthest along in this ecosystem, because of how much data they have. They can respond with much better responses. Siri is more general.

Skift: What do you think in terms of Facebook, and how far off are they from developing their own voice activated bot, such as a Siri or an Echo?

Preethi Kasireddy: I wouldn’t be surprised if they’re already working on it, honestly. I don’t see why not. Apple has one, Google has one, Amazon has one, Facebook seems to be the only one that doesn’t have one yet. Probably pretty soon. Think of how much data they have, and how many brand relationships they have, and all the advertising that’s happening on there. I wouldn’t be surprised they’re already doing it.

Skift: Just a follow-up question on the AI component of the chatbot ecosystem. You said that we haven’t really cracked the code on how to apply artificial intelligence and machine learning to the chatbot environment. Did I understand correctly, or is there significant progress?

You see a lot of that discussion kind of go hand-in-hand, where chatbots are intelligent, and the AI component allows the chatbot to learn, but it sounds like from you that we’ve not really made too much progress in that space.

Preethi Kasireddy: Yeah, AI is a really broad term. Right now, we are definitely using NLP, which is natural language processing. That is how the bot is able to understand the sense of what the user wants, so we are definitely using AI there, right now, even in our current bots.

From there, it’s somewhat manual. It’s basically an advanced if-else statement, and you have dialogs that flow through a bunch of results.

In terms of the bot being able to come up with its own dialogs and stuff like that, is where we haven’t gotten to. The AI of being able to respond to intent on its own based on the data we provide, is not something that we that we have gotten to yet, and it’s not able to get smarter.

Endnotes and Further Reading

  1. Vacation Rental Startups Attract $100 Million in Funding So Far This Year, Skift, May 2016.
  2. Facebook Bought Oculus For $3 Billion, Not $2 Billion: Zuckerberg Discloses Acquisition Details In Trial, Tech Times, January 2017
  3. Facebook Has a New Ad Product Tailor-Made for Travel Brands, Skift, September 2016
  4. Facebook Ads Bidding 101: Everything You Need to Know, AdEspresso, December 2016
  5. The US Travel Industry 2016: Digital Ad Spending Forecast and Trends, eMarketer, May 2016
  6. Priceline Buys Momondo for $550 Million to Expand in Europe, Skift, February 2017
  7. WhatsApp CEO on Facebook Deal: ‘It’s About Staying Independent’, The Wall Street Journal, February 2014
  8. WhatsApp Privacy Policy,
  9. Facebook Messenger now allows payments in its 30,000 chat bots, TechCrunch, September 2016
  10. The Most Popular Messaging App in Every Country, SimilarWeb, May 2016
  11. The amazing ways WeChat is used in China, Tech In Asia, December 2015
  12. Google Has a Messaging App Problem, The Wall Street Journal, January 2017
  13. Smartphone OS Market Share, 2016 Q3, IDC, October 2016
  14. Expedia Turns to Amazon Alexa for Its First Attempt at Voice Search, Skift, November 2016
  15. Microsoft Buys Skype for $8.5 Billion. Why, Exactly?, Wired, May 2011
  16. Skype has more than 300 million monthly active users, will get bots,, March 2016
  17. Line IPO Is Officially the Biggest American IPO This Year, Fortune, July 2016
  18. What’s Big in Japan?, SimilarWeb, March 2016
  19. Rakuten Drops on $900 Million Deal to Buy Viber Message App, Bloomberg, February 2014
  20. Viber Launching One-Touch Video Messaging To Stay Ahead Of Pack, PYMNTS.COM, December 2016
  21. Who is Telegram founder Pavel Durov?, CNN, December 2015
  22. Apple reportedly earning 0.15% from each Apple Pay transaction over iPhone 6/Apple Watch,, September 2014
  23.’s New Booking Messages Interface Empowers Customers to Personalize Their Stay in the Palm of Their Hand,, May 2016